Uploaded by Lishiyang1987

THE USE OF BEST MANAGEMENT PRACTICES (BMPS) IN URBAN WATERSHEDS

advertisement
United States
Environmental Protection
Agency
Office of Research and
Development
Washington DC 20460
EPA/600/R-04/184
September 2004
The Use of Best Management
Practices (BMPs) in Urban
Watersheds
EPA/600/R-04/184
September 2004
The Use of
Best Management Practices (BMPs)
in
Urban Watersheds
by
Swarna Muthukrishnan, Bethany Madge
Oak Ridge Institute for Science and Education
Edison, New Jersey 08837
and
Ari Selvakumar, Richard Field, Daniel Sullivan
U.S. Environmental Protection Agency
Edison, New Jersey 08837
Project Officer
Richard Field
Water Supply and Water Resources Division
National Risk Management Research Laboratory
U.S. Environmental Protection Agency
Edison, New Jersey 08837
NATIONAL RISK MANAGEMENT RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECION AGENCY
CINCINNATI, OHIO 45268
Foreword
The U.S. Environmental Protection Agency (EPA) is charged by Congress with protecting
the Nation's land, air, and water resources. Under a mandate of national environmental laws,
the Agency strives to formulate and implement actions leading to a compatible balance
between human activities and the ability of natural systems to support and nurture life. To
meet this mandate, EPA's research program is providing data and technical support for
solving environmental problems today and building a science knowledge base necessary to
manage our ecological resources wisely, understand how pollutants affect our health, and
prevent or reduce environmental risks in the future.
The National Risk Management Research Laboratory (NRMRL) is the Agency's center for
investigation of technological and management approaches for preventing and reducing risks
from pollution that threaten human health and the environment. The focus of the Laboratory's
research program is on methods and their cost-effectiveness for prevention and control of
pollution to air, land, water, and subsurface resources; protection of water quality in public
water systems; remediation of contaminated sites, sediments and ground water; prevention
and control of indoor air pollution; and restoration of ecosystems. NRMRL collaborates with
both public and private sector partners to foster technologies that reduce the cost of
compliance and to anticipate emerging problems. NRMRL's research provides solutions to
environmental problems by: developing and promoting technologies that protect and improve
the environment; advancing scientific and engineering information to support regulatory and
policy decisions; and providing the technical support and information transfer to ensure
implementation of environmental regulations and strategies at the national, state, and
community levels.
This publication has been produced as part of the Laboratory's strategic long-term research
plan. It is published and made available by EPA's Office of Research and Development to
assist the user community and to link researchers with their clients.
Lawrence W. Reiter, Acting Director.
National Risk Management Research Laboratory
-iii­
Disclaimer
Any opinions expressed in this paper are those of the author(s) and do not, necessarily, reflect
the official positions and policies of the U.S. EPA. Any mention of products or trade names
does not constitute recommendation for use by the U.S. EPA.
This document has been reviewed in accordance with the Environmental Protection
Agency’s peer and administrative review process, and it has been approved for publication
as an U.S. EPA document. It is a comprehensive literature review on commonly used urban
watershed Best Management Practices (BMPs) that heretofore was not consolidated. The
purpose of this document is to serve as an information source to individuals and
agencies/municipalies/watershed management groups/etc. on the existing state of BMPs in
urban stormwater management. Any information within the document should not be
considered as official U.S. EPA guidance. In addition, all reported recommendations or
values do not reflect the views of the agency and are solely owned by the cited sources.
-iv­
List of Select Acronyms
ASCE
BMP
BOD
Caltrans
Cd
COD
Cpv
CSO
Cu
CZARA
CZMA
DO
ED
EMAP
EMC
ENR
FC
FDEP
FHWA
ISMDSF
LID
MCTT
MEP
MS4
NGPE
NH3
NJDA
NO2NO3NPDES
NRCS
NURP
NVPDC
NYCDEP
NYSDEC
O&M
PAHs
Pb
American Society of Civil Engineers
best management practice
biochemical oxygen demand
California Department of Transportation
cadmium
chemical oxygen demand
channel protection storage volume
combined sewer overflow
copper
Coastal Zone Act Reauthorization Amendments
Coastal Zone Management Act
dissolved oxygen
extended-detention
Environmental Monitoring and Assessment Program
event mean concentration
Engineering News Record
fecal coliform
Florida Department of Environmental Protection
Federal Highway Administration
integrated stormwater decision-support framework
low-impact development
multi-chambered treatment train
maximum extent practicable
municipal separate storm sewer systems
natural growth protection easement
ammonia
New Jersey Department of Agriculture
nitrite
nitrate
National Pollutant Discharge Elimination System
Natural Resources Conservation Service
Nationwide Urban Runoff Program
Northern Virginia Planning District Commission
New York City Department of Environmental Protection
New York State Department of Environmental Conservation
operation and maintenance
polycyclic aromatic hydrocarbons
lead
-v­
PIT
QA/QC
QAPP
Rev
Rv
SS
SSC
SWPPP
TKN
TMDL
TP
TSS
U.S. EPA
USDA
USGS
WEF
WQv
WWF
Zn
pilot infiltration test
quality assurance/quality control
quality assurance project plan
recharge volume
runoff coefficient
suspended solids
suspended solids concentration
stormwater pollution prevention plan
total kjeldahl nitrogen
total maximum daily load
total phosphorus
total suspended solids
United States Environmental Protection Agency
United States Department of Agriculture
United States Geological Survey
Water Environment Federation
water quality volume
wet weather flow
zinc
-vi­
Contents
Chapter 1
The Use of Best Management Practices (BMPs) in Urban Watersheds ­
Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-1
Swarna Muthukrishnan, Bethany Madge, and Ari Selvakumar
Chapter 2
Types of Best Management Practices . . . . . . . . . . . . . . . . . . . . . . . . . . 2-1
Swarna Muthukrishnan, Richard Field, and Daniel Sullivan
Chapter 3
Structural BMP Design Practices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-1
Swarna Muthukrishnan, Richard Field, and Daniel Sullivan
Chapter 4
BMP Monitoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-1
Bethany Madge
Chapter 5
Effective Use of BMPs in Stormwater Management . . . . . . . . . . . . . . . 5-1
Bethany Madge
Chapter 6
BMP Cost . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-1
Ari Selvakumar
-vii­
1
The Use of Best Management Practices
(BMPs) in Urban Watersheds - Executive
Summary
Swarna Muthukrishnan, Bethany Madge, and Ari Selvakumar
CONTENTS
1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 1
1.2 The Concept of BMPs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 4
1.3 Summary and Findings of the White Paper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 6
1.4 Conclusions and Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 10
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 12
1.1 INTRODUCTION
Diffuse source pollution is one of the most intricate environmental problems with extensive
impacts on surface and groundwater quality. It is a major factor impacting the quality of water
supply, and the rate at which diffuse source pollutants are generated and delivered to water
resources is greatly affected by anthropogenic activities as well as natural processes. The main
hydrologic component transporting these pollutants to surface water bodies is runoff, which
results from precipitation or snowmelt (Leeds et al., 1993). Stormwater is part of a natural
hydrologic process; however, human activities, especially urban development and agriculture,
cause significant changes in patterns of stormwater flow from land into receiving waters. Urban
runoff can be or is often a significant source of water pollution, causing decline in fisheries,
swimming, and other beneficial attributes of water resources (U.S. EPA, 1993). Urban
stormwater runoff includes all flows discharged from urban land uses into stormwater
conveyance systems and receiving waters; in this context, urban runoff includes both dryweather non-stormwater sources (e.g., runoff from landscape irrigation, dewatering, and water
line and hydrant flushing) and wet-weather stormwater runoff. Water quality can also be affected
when runoff carries sediment and other pollutants into streams, wetlands, lakes, estuarine and
marine waters, or groundwater. The costs and impacts of water pollution from urban runoff are
significant and can include:
<
fish kills,
<
health concerns of human and/or terrestrial animals,
<
degraded drinking water,
<
diminished water-based recreation and tourism opportunities,
1-1
<
<
<
<
<
<
economic losses to commercial fishing and aquaculture industries,
lowered real estate values,
damage to habitat of fish and other aquatic organisms,
inevitable costs of clean-up and pollution reduction,
reduced aesthetic values of lakes, streams, and coastal areas, and
other impacts (Leeds et al., 1993).
Increased stormwater flows from urbanization have the following major impacts (FLOW, 2003):
<
acceleration of stream velocities and degradation of stream channels,
<
declining water quality due to washing off of accumulated pollutants from impervious
surfaces to local waterways, and an increase in siltation and erosion of soils from
pervious areas subject to increased runoff,
<
increase in volume of runoff with higher pollutant concentrations that reduces receiving
water dilution effects,
<
diminished groundwater recharge, resulting in decreased dry-weather flows; poorer water
quality of streams during low flows; increased stream temperatures; and, greater annual
pollutant load delivery,
<
increased flooding,
<
combined and sanitary sewer overflows due to stormwater infiltration and inflow,
<
damage to stream and aquatic life resulting from suspended solids accumulation, and
<
increased health risks to humans from trash and debris which can also endanger and
destroy food sources or habitats of aquatic life (FLOW, 2003).
The major categories of stormwater pollutants, their sources, and related impacts are presented in
Table 1-1.
Table 1-1. Major Categories of Stormwater Pollutants, Sources and Related Impacts
Stormwater Pollutant
Major Sources
Related Impacts
Nutrients:
Nitrogen, Phosphorus
Urban runoff; failing septic
systems; croplands; nurseries;
orchards; livestock
operations; gardens; lawns;
forests; fertilizers;
construction soil losses
Algal growth; reduced
clarity; lower dissolved
oxygen; release of other
pollutants; visual
impairment; recreational
impacts; water supply
impairment
1-2
Solids:
Sediment (clean and
contaminated)
Construction sites; other
disturbed and/or nonvegetated lands; road
sanding; urban runoff;
mining operations; logging
operations; streambank and
shoreline erosion
Increased turbidity; reduced
clarity; lower dissolved
oxygen; deposition of
sediments; smothering of
aquatic habitat including
spawning sites; sediment and
benthic toxicity
Oxygen-depleting substances
Biodegradable organic
material such as plant; fish;
animal matter; leaves; lawn
clippings; sewage; manure;
shellfish processing waste;
milk solids; other food
processing wastes;
antifreeze/other de-icing
chemicals; other applied
chemicals
Suffocation or stress of adult
fish, resulting in fish kills;
reduction in fish reproduction
by suffocation/stress of
sensitive eggs and larvae;
aquatic larvae kills; increased
anaerobic bacterial activity
resulting in noxious gases or
foul odors often associated
with polluted water bodies;
release of particulate bound
pollutants
Pathogens:
Bacteria, Viruses, Protozoans
Domestic and natural animal
wastes; urban runoff; failing
septic systems; landfills;
illegal cross-connections to
sanitary sewers; natural
generation
Human health risks via
drinking water supplies;
contaminated shellfish
growing areas and swimming
beaches; incidental ingestion
or contact
Metals:
Lead, Copper, Cadmium,
Zinc, Mercury, Chromium,
Aluminum, others
Industrial processes; mining
operations; normal wear of
automobile brake pads and
tires; automobile emissions;
automobile fluid leaks; metal
roofs; gutters; landfills;
corrosion; urban runoff; soil
erosion; atmospheric
deposition; contaminated
soils
Toxicity of water column and
sediment; bioaccumulation in
aquatic species and through
food chain
1-3
Hydrocarbons:
Oil and Grease, Polyaromatic
hydrocarbons (PAHs) - e.g.,
Naphthalenes, Pyrenes
Industrial processes;
automobile wear; automobile
emissions; automobile fluid
leaks; waste oil
Toxicity of water column and
sediment; bioaccumulation in
aquatic species and through
food chain; lower dissolved
oxygen (DO); coating of
aquatic organism gills/impact
on respiration
Organics:
Pesticides, Polychlorinated
biphenyls (PCBs), Synthetic
chemicals
Applied pesticides
(herbicides, insecticides,
fungicides, rodenticides,
etc.); industrial processes;
nurseries; orchards; lawns;
gardens; historically
contaminated soils/wash-off
Toxicity of water column and
sediment; bioaccumulation in
aquatic species and through
food chain
Toxicity of water column and
Irrigated lands; mining
sediment
operations; landfills; road
salting and uncovered salt
storage
(Leeds et al., 1993; MA DEP and MA CZM, 1997; U.S. EPA, 2000)
Inorganic Acids and Salts
(sulphuric acid, sodium
chloride)
1.2 THE CONCEPT OF BMPS
The undesirable impacts of stormwater runoff can be controlled by prudent management efforts.
Stormwater management encompasses an array of measures that involve careful application of
site design principles, construction techniques to prevent sediments and other pollutants from
being released and/or entering surface or groundwater, source controls, and treatment of runoff to
reduce pollutants and reducing the impact of altered hydrology.
For many years, federal and state regulations for stormwater management efforts were oriented
towards flood control with minimum measures directed towards improving the quality of
stormwater such as sediments and erosion control and the reduction of pollutants. The U.S.
recognized the problem of diffuse pollution many years ago and established provisions in a major
amendment to the Clean Water Act in 1987, leading to national programs of action to address the
issue. The increased awareness of the need to improve water quality in the last two decades
resulted in the concept of best management practices (BMPs) which are measures intended to
provide an on-the-ground practical solution to diffuse pollution problems from all sources and
sectors (D'Arcy and Frost, 2001). BMPs are technology and education based requirements in the
federal stormwater regulations that call for the implementation of controls to reduce the
discharge of pollutants to the Maximum Extent Practicable (MEP) in municipal-type stormwater
systems (Caltrans, 2002). BMP refers to operational activities, physical controls or educational
1-4
measures that are applied to reduce the discharge of pollutants and minimize potential impacts
upon receiving waters, and accordingly, refers to both structural and nonstructural practices that
have direct impacts on the release, transport, or discharge of pollutants.
The proper management of stormwater runoff is necessary to reduce stream channel erosion,
pollution, siltation, sedimentation, and local flooding, all of which have adverse impacts on the
land, water resources, and the people. The BMP program was increasingly designed in the 1980s
primarily to address pollution from wet-weather flow (WWF) and polluted runoff and focused on
controlling runoff increases and reducing water quality degradation associated with new
development. The goal of these practices is to maintain the predevelopment characteristics as
close as possible, even after development of a site, and/or to reduce the impacts to an accepted
level. It must be understood in this context that BMPs do not merely act as controls for new
development, but these practices equally apply to existing developments as well as areas that
have undergone any kind of re-development.
The BMP concept has the following key elements (D'Arcy and Frost, 2001):
<
There is a need for guidance that offers practical prevention options.
<
The options need to be defined and explicit best practice rather than ill-defined individual
interpretations of what is required.
<
The options should be describable as best practice, based on research and experience.
Since the development of BMPs, various state and local governments have adopted a profusion
of laws, regulations, and policies to encourage or mandate the use of urban BMPs. These BMPs
have been developed and refined to mitigate some, if not all, of the adverse impacts associated
with any kind of development/re-development activity. The capabilities of each BMP are
unique. This needs to be recognized along with its limitations, and these factors, in addition to
the physical constraints at the site, need to be judiciously balanced with the overall management
objectives for the watershed in question. At a minimum, a BMP program developed for a site
should strive to accomplish the following set of criteria:
<
Reproduce, as nearly as possible, the natural hydrological conditions in the stream prior
to development or any previous human alteration (Schueler, 1987; Young et al., 1996).
<
Provide a moderate to high level of removal for most urban pollutants as one of a set of
BMPs in the watershed working together to achieve desired receiving-water quality.
<
Be appropriate for the site, given physical constraints.
<
Be reasonably cost-effective in comparison with other BMPs.
<
Have an acceptable future maintenance burden.
<
Have a neutral impact on the natural and human environment.
The purpose of this white paper is to provide a general description and insight on the various
BMP options, the design considerations involved and the general guidelines for selection,
implementation, and monitoring of BMPs to reduce pollutants in urban stormwater from new
development and re-development. As the main focus of this white paper is structural BMPs, the
various nonstructural practices is discussed only briefly. This white paper however, does not
1-5
intend to dictate or specify the actual selection of BMPs, but attempts to provide the framework
for an informed selection of BMPs for any stormwater management program.
1.3 SUMMARY AND FINDINGS OF THE WHITE PAPER
Chapter 2 provides a general discussion of the most commonly used nonstructural and structural
BMPs for the management of urban storm runoff. The introduction defines what a BMP is and
describes what structural and nonstructural BMPs are. The chapter discusses how the distinction
between structural and nonstructural BMPs is quite clear in many cases, and not entirely so in the
case of some other BMPs. Some good examples related to emerging concepts in runoff
management are presented, such as better site design and Low-Impact Development (LID)
techniques e.g., green roofs (Section 2.1), that focus on the use of both site planning and smallscale treatment approaches. These practices reinforce the growing opinion that a combination of
both structural and nonstructural practices in a treatment train approach is almost certainly a
better option to meet stormwater management objectives for many project sites. Section 2.2
briefly describes the types of structural BMPs currently being used and recommended by the U.S.
EPA's menu of BMPs (U.S. EPA, 2001) and several state agencies. The general advantages and
limitations of each of these BMPs are also discussed. Section 2.3 focuses on the major
physicochemical and biological processes in BMPs and how they influence the removal of
pollutants and mitigate other stressors from stormwater, and section 2.4 presents a brief overview
of the factors influencing the performance of these structural BMPs.
In Chapter 3, Structural BMP Design Practices, the factors that need to be considered in
designing urban BMPs are discussed. The following eight BMPs commonly used for stormwater
treatment in new development are addressed: (i) dry extended-detention ponds; (ii) wet ponds;
(iii) stormwater wetlands; (iv) grassed swales; (v) vegetated filter strips; (vi) infiltration trenches;
(vii) porous pavement; and, (viii) sand and organic filters. As an introduction, Section 3.1 traces
the development of the concept of BMPs from earlier stormwater management measures that first
relied heavily on flood and then water quantity control to the current focus on controlling both
the quality and quantity of runoff in order to mitigate the impacts to receiving waters. Section
3.2 briefly explains the sizing criteria involved in BMP design considerations and describes the
performance objectives of these BMPs. Section 3.3 describes the design considerations for the
above-mentioned eight BMPs in detail in separate subsections 3.3.1 - 3.3.8. Each subsection on
a specific BMP carries a brief description of the BMP and how it addresses the two issues of
stormwater control as well as pollutant removal. Each subsection also has a detailed presentation
on the general design considerations including, site suitability, physical specifications, and
geometry.
The design and construction of stormwater BMPs is a constantly evolving process in that there
does not appear to exist a "100 % fool-proof" design for a single BMP that can achieve the entire
spectrum of desirable stormwater benefits in a watershed. A clear understanding of the key
mechanisms within a BMP for effluent load reductions and factors that govern these processes is
1-6
a primary requirement in designing a BMP. On the same note, it must be mentioned that there
has been no "exact" or "perfect" design to date. Performance variations in BMPs discussed in
Chapter 2 is a fallout of these differences in design characteristics. Chapter 3 identifies the
following key issues which need to be addressed in order to improve the design of stormwater
BMPs:
<
Influent mass loadings should be defined more clearly by considering all associated
parameters that include flow rate, pollutant concentrations and their chemical forms,
suspended solids and their settling velocities, dissolved solids, and the size apportionment
of pollutants in the solid phase.
<
Approaches to designing BMPs should focus on frequently-occurring smaller storms; the
focus should be on characterizing influent load in such smaller storms and especially the
parameters of concern in each watershed.
<
BMP design should integrate engineering principles with hydrological characteristics,
BMP performance objectives, flow attenuation, and flood control, and should incorporate
design features that would enhance the BMP capability in treating the stressor(s) of
concern.
<
Designers should realize that one BMPs is not adequate to address the above mentioned
issues and should consider the use of a treatment train; i.e., a combination of structural
and nonstructural BMPs in stormwater treatment programs.
BMP Monitoring, Chapter 4, covers the complexities in developing a BMP monitoring program
that yields useful results. Some difficulties and criticisms with current BMP monitoring
practices are discussed in Section 4.1. This section identifies many key areas of BMP monitoring
programs that require improvement in order for the data of such programs to be widely
applicable. Sections 4.2 through 4.5 contain recommendations and explanations of how to
ameliorate these deficiencies. Selection of appropriate parameters is covered in Section 4.2.
Considerations and difficulties with monitoring nonstructural BMPs and watersheds as a whole
are presented in Sections 4.3 and 4.4, respectively. Guidance on the development of a robust
structural BMP effectiveness monitoring program is presented in Section 4.5. This section
includes information and recommendations on planning, designing, implementing, and
evaluating BMP monitoring programs. Methods for data analysis, which is particularly
inconsistent between studies, is discussed in subsection 4.5.4. After reading this section, the
reader should have a good handle on the problems and complexities associated with BMP
effectiveness monitoring programs. The reader should also be equipped with the general
knowledge and understanding necessary to develop a successful BMP monitoring program,
complete with representative, quality assured, and statistically analyzed results.
BMP monitoring, especially for effectiveness, is a very complex undertaking. The number of
variables that affect the resulting efficiency of a BMP is large. This, along with nonuniform
sampling and analysis techniques used in current monitoring programs, has led to a wide degree
of variability in reported BMP performances. The selection of appropriate pollutants is one of
the most fundamental requirements in a robust BMP monitoring program. Due to the large
number of variables involved in BMP performance, selecting a reasonable number of suitable
1-7
parameters is difficult and requires experience and good guidance. Nonstructural BMPs have
even more complications with respect to monitoring effectiveness. Without a defined influent
and effluent, monitoring programs usually rely on public surveys or watershed monitoring
approaches. Watershed monitoring approaches, while seemingly economical when a large
number of BMPs require monitoring, is wrought with interferences from outside sources
unrelated to BMPs, such as intrusion of contaminated groundwater or the inability to distinguish
individual BMP performance. Thus, data from these types of monitoring programs may not be
able to produce the results that shed light on the effectiveness of BMPs within the watershed.
Four steps have been outlined to monitor the effectiveness of structural BMPs. This guideline
assists in the development of a robust monitoring program from planning and design phases,
through the implementation and evaluation phases. A BMP monitoring program should always
be initiated with clear goals and specific objectives backed by supporting background
information. This foundation will minimize the risk of collecting data that is not useful. Once
the goals and objectives are identified, a quality assurance project plan (QAPP) translates
objectives into a plan of action. Producing a useful QAPP requires a significant amount of time
upfront before any samples are taken. Although it may seem at first to be a tedious exercise, it
will likely save time and money in the long run by ensuring the significance of the data collected.
Design aspects, such as monitoring approach, parameter and methods selection, specifics on
hydraulic, hydrologic, and water quality data collection, and methods of analysis, equipment
selection, and quality assurance/quality control measures should all be clearly stated in the
QAPP. Hydrologic and hydraulic (H&H) data is one of the most essential components of a well
designed monitoring program. Poor quality H&H data will produce errors that will propagate
through the rest of the results. Whether the QAPP calls for composite or discrete samples, they
should always be flow-weighted or synchronized with flow measurements in some way.
Representativeness of the collected samples is another key component that is often overlooked
in BMP monitoring programs. Guidelines such as percent capture and minimum number of
aliquots should be used to ensure each storm event is accurately represented. Once samples are
collected, the chosen method of data analysis must provide useful and unbiased results. For
example, percent removal is biased against BMPs with relatively clean influent and may not be
useful for watersheds with very high influent loads as the resulting effluent, even with high
percent removals, would not meet overall water quality objectives; similarly, any method that
produces pollutant concentrations instead of loads will be biased against BMPs that rely on
infiltration.
When BMPs are used in stormwater management, many issues need to be addressed to ensure
that the BMPs are being used as effectively as possible. Chapter 5, Effective use of BMPs in
Stormwater Management, covers these issues. The proper selection of a BMP is one key
component to the effective use of BMPs in stormwater management. One must consider
regulatory constraints, site factors, the ability of the BMP to provide stormwater quality and
quantity control, cost, reliability, maintenance burden, and environmental and community
acceptance. Section 5.1 walks through all these considerations and provides useful information
in tabular format on many of the commonly used BMPs. Structural BMP placement is currently
1-8
a “hot-button” issue. Key concerns regarding optimum and appropriate placement options are
covered in section 5.2. The chapter is rounded out by a discussion of BMP integration.
An integrated approach to stormwater management appears to be the most effective use of
BMPs. When multiple layers of structural and nonstructural BMPs are used in unison, the
watershed will reap the largest benefit. The selection of BMPs to be used within such an
integrated approach (or as single units) is dependant primarily on applicable regulations and
estimated water quality and quantity performance. However, it is once again stressed that instead
of relying solely on numerical efficiencies reported in the literature, a much deeper understanding
of the factors that control BMP pollutant removal performance is essential to proper selection
and design of a BMP. Other factors such as site characteristics, cost, reliability, maintenance
requirements, and environmental and community acceptance also need consideration to ensure
the chosen BMP performs as desired and expected.
BMP placement is a relatively new issue in stormwater management. Currently, political issues
such as regulations that require BMPs for approval of new construction permits often control
BMP placement. For this reason onsite placement is the only option. Although, onsite
placement of BMPs has its advantages, more uniform sub-regional and regional BMP placement
have their advantages as well. Recent efforts to identify optimal BMP placement in watersheds
through modeling efforts may uphold or challenge the current focus on onsite placement
practices.
Chapter 6 provides information on how to estimate the cost of structural BMPs. Costs of
nonstructural BMPs are not included here as they are generally not as easily quantified as
structural BMPs due to their indirect nature. Section 6.1 covers BMP cost estimating procedures
and discusses four common methods of estimating costs. Section 6.2 contains information on
total costs which include both capital (construction and land) and annual operation and
maintenance costs. BMPs can present several tangible economic benefits in spite of their high
construction (in certain cases), operation, and maintenance costs and these are discussed in
Section 6.3. At other times, the use of BMPs can result in reduced infrastructure costs.
The cost of constructing any BMP is variable and can be substantial. The cost of constructing a
BMP depends on many factors, including the time of year, site conditions and topography,
accessibility of equipment, economics of scale, and government regulations. Several documents
have been published that address cost estimation for BMPs, but most of these report only
construction costs (Young et al., 1996; Sample et al., 2003). In addition, costs are often
documented as base costs and do not include land costs, which is the largest variable influencing
overall BMP cost (U.S. EPA, 1999). However, in some areas with minimum landscaping
requirements, the implementation of standard practices may mean there are no “extra” land costs.
The wide range of cost data reported in the literature indicates that much more information is
needed in this area.
1-9
1.4 CONCLUSIONS AND RECOMMENDATIONS
The use of BMPs to control and treat urban stormwater runoff has become a common practice in
urban watershed management. This has been propagated by ordinances developed by local
governments that dictate the use of structural and nonstructural BMPs for new and existing
development and to protect surface water quality and mitigate the impacts of stormwater runoff
on receiving waters. BMPs demonstrate a wide range of pollutant removal capabilities and their
performance is affected by several factors, including the long-term variation in rainfall, BMP
design characteristics, processes affecting chemical phase and speciation, and environmental
conditions. The pollutant removal performance of BMPs is difficult to interpret beyond
generalities due to various inter-related and complex parameters; shortcomings in current BMP
related studies include:
<
lack of long-term monitoring of the processes in a BMP responsible for export or
detention/retention of urban stormwater pollutants,
<
absence of, or inadequate monitoring within a BMP for water quality, sediment, and
vegetation, which would provide a strong understanding of factors and processes that
affect pollutant fate within a BMP, and
<
variability in BMP performance results not only due to factors affecting the performance
of BMPs, but also the methods used to characterize and calculate BMP effectiveness.
Present and future research initiatives on the implementation of structural BMPs for effective
stormwater management resulting in water quality and quantity control should address the
following issues:
<
BMPs can be more effectively managed with a better understanding of the specific
physico-chemical and biological processes and interactions that govern the
transformation, immobilization and export of pollutants in stormwater. A national
approach, similar to the Nationwide Urban Runoff Program (NURP), which would
systemize a large number of investigations into a cohesive, well-controlled, program to
learn about various BMP functions, physical mechanisms, biochemistry, and design
parameters, is very much needed and in fact is underway (Urbonas, 2000). Also, the
interactions among these various processes need to be understood. A knowledge of
internal dynamics within a BMP, such as a pond or a wetland, is essential to improve
upon BMP design and maintenance considerations.
<
The type of pollutants studied should include a more extensive array of priority pollutants
such as hydrocarbons, toxic inorganics and pathogens. There is less monitoring data
available for pollutants such as dissolved and particulate metal species, hydrocarbons, and
bacteria, (U.S. EPA, 1999). Bacteria and viruses are rarely sampled in stormwater
studies, but a number of microbial pathogens can be present in stormwater and have been
implicated in waterborne disease outbreaks in both humans and fish populations
(Rushton, 2002). There is a need to improve existing analytical tools and establish new
metrics to detect and estimate the concentrations of these pollutants. Two important
1 - 10
research priorities include: (i) the need to develop comprehensive methods to assess the
heavy metal bioavailability and other toxics following BMP treatment; and, (ii) the
development of meaningful metrics to study stormwater pathogens and their ecological
and health impacts. The use of conventional means of assessing water quality by only
estimating the fecal coliform (FC) count is not applicable to stormwater (O'Shea and
Field, 1992; Rushton, 2002). Stormwater has a wider array of pathogens including
bacteria, fungi, viruses, and protozoans such as Cryptosporidium and Giardia that are not
well characterized by existing methods including the FC test.
<
With regard to monitoring programs, commonly used BMPs such as infiltration trenches,
infiltration basins, bioretention practices, and filter strips are seldom monitored (U.S.
EPA, 1999). This could be due to the difficulties associated with collecting inflow and
outflow samples from these systems. These systems are widely used, yet their
effectiveness has not been well documented; pilot-scale research investigations would be
a good approach to assess their effectiveness.
<
BMP monitoring should be conducted over a relatively long time period (one yr or more)
continuously during dry- and wet-weather flow conditions using an influent-effluent mass
balance approach. BMP monitoring and data reporting should be more frequent, (e.g., on
a monthly and seasonal basis) rather than estimating annual averages of pollutant
removal. This temporal scale approach helps to better assess seasonal factors such as
thermal stratification of ponds, plant growth and senescence that influence the
performance of BMPs. Also, the investigation of dry- weather/low- flow samples would
be the key to ascertain if pollutants are released under low flow conditions.
<
The current approach in the use of urban BMPs perceives the effect these practices would
have just on water quality. BMPs are best characterized by: (i) how much runoff is
prevented; (ii) volume of runoff being treated; and, (iii) the quality of the resulting
effluent. In addition, BMP objectives should foremost extend beyond water quality and
take into account other media including sediments, vegetation, and benthic invertebrates.
The role of vegetation, bacteria, and benthic invertebrates in the accumulation and/or
breakdown of pollutants need to be more intensively investigated.
<
Most of the commonly used structural BMPs extensively rely on sedimentation as the
predominant mechanism to remove pollutants from the water column. What is largely
ignored is that these contaminated sediments should be subject to appropriate
maintenance practices; they have the potential to either leach the organic and inorganic
contaminants into the ground below, or may result in pollutant resuspension during high
flows or under unfavorable environmental conditions, which include changes in pH
and/or the oxidation reduction potential (ORP). Any BMP research or monitoring
program should have in its analytical scheme the sediment speciation for contaminant
assessment in order to understand the bioavailability and the potential for pollutant
resuspension in the water column.
1 - 11
<
A large number of BMP studies at present focus on seasonal, short-term monitoring
activities; the long-term performance of BMPs (>5 yr) is uncertain. The pollutant
removal capabilities of BMPs are likely limited by a finite capacity of sediment/substrate
to sorb and retain pollutants. Numerous research findings infer that the longevity of
BMPs is linked to the ability of the substrate to assimilate pollutants and maintenance
practices (Schueler et al., 1992), lending credence to the need for an evaluation of the
long-term performance of these systems.
REFERENCES
State of California Department of Transportation (Caltrans). (2002). Stormwater Quality
Handbooks Project Planning and Design Guide.
http://www.dot.ca.gov/hq/oppd/stormwtr/PPDG-stormwater-2002.pdf.
D'Arcy, B. and A. Frost. (2001). The role of best management practices in alleviating water
quality problems associated with diffuse pollution. The Science of The Total Environment
265: 359-367.
Friends of the Lower Olentangy Watershed (FLOW). (2003). The Lower Olentangy Watershed
Action Plan in 2003 - Strategies for Protecting and Improving Water Quality and
Recreational Use of the Olentangy River and Tributary Streams in Delaware and
Franklin Counties. Columbus, OH.
http://www.olentangywatershed.org/LowerOlentangyActionPlan04.pdf.
Leeds, R., L.C. Brown, and N.L. Watermeier. (1993). Nonpoint Source Pollution: Water Primer.
AEX-465-93. Ohio State University Extension Fact Sheet. Food, Agricultural and
Biological Engineering, The State University of Ohio, Columbus, OH.
http://ohioline.osu.edu/aex-fact/0465.html
Massachussets Department of Environmental Protection and Massachussets Office of Coastal
Zone Management (MA DEP and MA CZM). (1997). Stormwater Management Volume
Two: Stormwater Technical Handbook. Boston, MA.
Rushton, B. (2002). BMP monitoring: methods and evaluations, http://www.stormwaterresources.com/library.htm#BMPs (#147PL). Stormwater News, Cocoa Beach, Florida.
Sample, D.J., J.P. Heaney, L.T. Wright, C.Y. Fan, F.H. Lai, and R. Field. (2003). Cost of best
management practices and associated land for urban stormwater control. Journal of Water
Resources Planning and Management 129(1): 59-68.
Schueler, T.R. (1987). Controlling Urban Runoff: A practical manual for planning and
designing urban BMPs. Metropolitan Washington Council of Governments, Washington,
DC.
Schueler, T.R., P.A. Kumble, and M.A. Hearty. (1992). A Current Assessment of Urban Best
Management Practices: Techniques for reducing nonpoint source pollution in the coastal
zone. Metropolitan Washington Council of Governments, Washington, DC.
United States Environmental Protection Agency (U.S. EPA). (1993). Handbook - Urban Runoff
Pollution Prevention and Control Planning. EPA-625-R-93-004. Washington, DC.
United States Environmental Protection Agency (U.S. EPA). (1999). Preliminary Data Summary
1 - 12
of Urban Stormwater Best Management Practices. EPA-821-R-99-012. Washington, DC.
United States Environmental Protection Agency (U.S. EPA). (2000). National Water Quality
Inventory 1998 Report to Congress. EPA-841-R-00-001. Washington, DC.
United States Environmental Protection Agency (U. S. EPA). (2001). National Menu of Best
Management Practices for Stormwater Phase II,
http://cfpub.epa.gov/npdes/stormwater/menuofbmps/menu.cfm. Washington, DC.
Urbonas, B. (2000). Assessment of stormwater best management practice effectiveness (chapter
7). In: Innovative urban wet-weather flow management systems, Field, R., J.P. Heaney,
and R. Pitt, (editors). Technomic Publishing Co., Inc., Lancaster, PA. Pages 255 - 300.
Young, G.K., S. Stein, P. Cole, T. Kammer, F. Graziano, and F. Bank. (1996). Evaluation and
Management of Highway Runoff: Water Quality Technical Report. Department of
Environmental Programs, Metropolitan Washington Council of Governments,
Washington, DC.
1 - 13
2
Types of Best Management Practices
Swarna Muthukrishnan, Richard Field and Daniel Sullivan
CONTENTS
2.0 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 2
2.1 Nonstructural or Source Control BMPs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 4
2.1.1 Public Education . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 6
2.1.2 Planning and Management of Developing Areas . . . . . . . . . . . . . . . . . . . . . . . . 2 - 7
2.1.2.1 Better Site Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 7
2.1.2.2 Vegetation Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 8
2.1.2.3 Reduction/Disconnection of Impervious Areas . . . . . . . . . . . . . . . . . . 2 - 8
2.1.2.4 Green Roofs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 8
2.1.2.5 Low-Impact Development (LID) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 10
2.1.3 Materials Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 11
2.1.3.1 Alternative Product Substitution . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 11
2.1.3.2 Housekeeping Practices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 12
2.1.4 Street/Storm Drain Maintenance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 13
2.1.4.1 Street Cleaning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 13
2.1.4.2 Catchbasin Cleaning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 14
2.1.4.3 Roadway and Bridge Maintenance . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 14
2.1.4.4 Storm Drain Flushing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 15
2.1.4.5 BMP Maintenance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 15
2.1.4.6 Storm Channel and Creek Maintenance . . . . . . . . . . . . . . . . . . . . . . . 2 - 16
2.1.4.7 Stormwater “Hotspots” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 16
2.1.5 Spill Prevention and Cleanup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 17
2.1.5.1 Above Ground Tank Spill Control . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 17
2.1.5.2 Vehicle Spill Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 17
2.1.6 Illegal Dumping Controls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 17
2.1.6.1 Storm Drain Stenciling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 18
2.1.6.2 Household Hazardous Waste Collection . . . . . . . . . . . . . . . . . . . . . . 2 - 18
2.1.6.3 Used Oil Recycling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 19
2.1.7 Illicit Connection Controls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 19
2.1.7.1 Illicit Connection - Prevention, Detection and Removal . . . . . . . . . . 2 - 20
2.1.7.2 Failing Septic Systems and Sanitary Sewer Overflows . . . . . . . . . . . 2 - 20
2.1.8 Stormwater Reuse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 21
2.2 Structural or Treatment Best Management Practices . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 21
2.2.1 Stormwater Ponds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 22
2-1
2.2.1.1 Dry Detention Ponds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 22
2.2.1.2 Dry, Extended-detention (ED) Ponds . . . . . . . . . . . . . . . . . . . . . . . . 2 - 23
2.2.1.3 Wet (Retention) Ponds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 24
2.2.2. Stormwater Wetlands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 25
2.2.2.1 Constructed Wetlands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 25
2.2.3 Vegetated Systems (Biofilters) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 26
2.2.3.1 Grass Swales (Wet, Dry) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 26
2.2.3.1.1 Dry Swales . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 27
2.2.3.1.2 Wet Swale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 27
2.2.3.2 Vegetative Filter Strips (VFS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 28
2.2.3.3 Bioretention . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 29
2.2.4 Infiltration Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 31
2.2.4.1 Infiltration Basins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 32
2.2.4.2 Infiltration Trenches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 33
2.2.4.3 Porous Pavement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 34
2.2.5 Filtration Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 35
2.2.5.1 Surface Sand Filter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 36
2.2.5.2 Underground Vault Sand Filter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 36
2.2.5.3 Perimeter Filter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 36
2.2.5.4 Others . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 37
2.2.6 Technology Options and Others . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 37
2.2.6.1 Water Quality Inlets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 38
2.2.6.2 Multi Chambered Treatment Train (MCTT) . . . . . . . . . . . . . . . . . . 2 - 39
2.2.6.3 Vortex Separators and Continuous Deflective Systems . . . . . . . . . . 2 - 40
2.3 Removal Processes Occurring in Structural BMPs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 42
2.3.1 Sedimentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 43
2.3.2 Filtration
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 45
2.3.3 Sorption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 45
2.3.4 Phytoremediation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 46
2.3.5 Biological Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 47
2.4 Performance of Structural BMPs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 48
2.4.1 Suspended Solids, Nutrients, and Heavy Metals, and BMP Performance . . . . 2 - 50
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 52
2.0 INTRODUCTION
A stormwater best management practice (BMP) is a technique, measure, or structural control that
is used to manage the quantity and improve the quality of stormwater runoff in the most cost
effective manner. The U.S. EPA (1999b) defines best management practices as "schedules of
activities, prohibitions of practices, maintenance procedures, and other management practices to
prevent or reduce the pollution of waters of the United States.” BMPs also include, but are not
limited to, treatment requirements, operating procedures, and practices to control plant site
2-2
runoff, spillage or leaks, sludge or wastewater disposal, or drainage from raw material storage."
There are two general types of BMPs used to reduce the threat of stormwater runoff pollution
from construction and development in urbanizing areas: (i) Nonstructural or source control
BMPs and (ii) Structural or treatment BMPs (U.S. EPA, 1993b).
Nonstructural BMPs refer to those stormwater runoff management techniques that use natural
measures to reduce pollution levels, do not require extensive construction efforts, and either limit
the generation of stormwater runoff, or reduce the amounts of pollutants contained in the runoff.
They do not involve fixed, permanent facilities and they usually work by changing behavior
through government regulation (e.g., planning and environmental laws), persuasion, and/or
economic instruments (Taylor and Wong, 2002). These BMPs include institutional, educational
or pollution prevention practices. Because they improve runoff quality by reducing the use,
generation and accumulation of potential stormwater contaminants at or near their sources in
many cases, they are also termed as source control BMPs (WEF and ASCE, 1998; U.S. EPA,
1999b).
Structural BMPs are engineered systems and methods designed to provide temporary storage and
treatment of stormwater runoff for the removal of pollutants (MWLAP, 1992; MDE, 2000; Clar
et al., 2003). These practices are aimed at "controlling the volume and discharge rate of
stormwater runoff from urban areas, as well as, reducing the magnitude of pollutants in the
discharge water through rotation, physical containment or flow restrictions designed to allow
settling, physical removal through filtration, percolation, chemical precipitation or flocculation,
and/or biological uptake and may be called structural or treatment BMPs" (Florida DER, 1988).
Structural BMPs improve the quality and/or control the quantity of stormwater runoff; common
examples include detention ponds and constructed wetlands.
The distinction between nonstructural and structural BMPs is very clear in some cases, but less
so in others. For example, street sweeping for pollutant removal is one BMP that could be
considered either a structural or nonstructural control (WEF and ASCE, 1998). The use of
vegetation to disconnect directly impervious surfaces such as rooftops, driveways, parking lots
and streets, is another example of a BMP that could be considered as a structural or nonstructural
control. Some of the newer concepts for urban stormwater runoff management such as better site
planning techniques (CWP, 1998) and low-impact development (LID) technologies such as
"green roofs" (U.S. EPA, 2000a, 2000b) focus on the use of both planning techniques and
micro-scale integrated landscape-based practices to prevent or reduce the impacts of urban
stormwater runoff at the very point where these impacts would be initially generated or just
downstream. These approaches tend to have a great deal of overlap between preventative source
control approaches and small-scale treatment approaches that blur the distinction between these
two types of BMPs. It is being recognized that a combination of nonstructural controls and one
or more structural or treatment BMPs, often referred to as a "treatment train approach," may be
needed to meet water quality objectives, depending on the stormwater management goals and
objectives identified for a specific site or area. For example, if the pollutants of concern are a
combination of trash/debris, dissolved copper and phosphorus, it is more likely that a treatment
2-3
train would be required to be successful in treating the stormwater runoff.
The list of BMPs, both structural and nonstructural, presented in the following discussion is
based on information available at the time that this white paper was written. The development of
stormwater BMPs is an emerging science and stormwater treatment technologies are frequently
being modified; also, efforts are increasingly directed towards pollution prevention and the
implementation of nonstructural practices. Finally it is being increasingly recognized that given
the number and forms of the pollutants of concern, multiple treatment processes are required.
Keeping this in mind, the list of BMPs is provided in this document as a guide only and is not
meant to be exclusive of alternative stormwater BMPs that could be used in urban watershed
management.
2.1 NONSTRUCTURAL OR SOURCE CONTROL BMPS
Nonstructural or source control BMPs are practices that prevent pollution by reducing potential
pollutants at their source before they come into contact with stormwater, or capturing and
disposing of stormwater at its source. These BMPs aim to eliminate contamination by preventing
their introduction into the environment. Often termed pollution prevention practices, these
BMPs can include any method that avoids or reduces potential exposure of contamination to the
elements. Unfortunately, nonstructural BMPs are often not being given due credit in watershed
management plans (Clar et al., 2003) as data on their performance is virtually not available.
However, there is an increasing recognition of the primary need for pollution prevention rather
than treatment of polluted stormwater in long-term urban watershed management program
implementation. There are two major reasons for this shift in the approach to stormwater
management:
<
<
Stormwater management approaches are increasingly being directed toward the
development of least-cost measures to treat stormwater pollution, which include
nonstructural and low-cost structural controls.
Nonstructural BMPs are found to be very effective in controlling pollution generation at
the source and can eliminate or reduce the need for costly end-of-pipe treatment by
structural BMPs (WEF and ASCE, 1998).
Nonstructural controls include regulatory controls that prevent pollution problems by controlling
land development and land use, as well as source controls that reduce pollutant buildup or lessen
its availability for washoff during rainfall (City of Austin, 1988; MWLAP, 1992; U.S. EPA,
2000a, 2000b). Nonstructural practices can play a significant role in reducing water quality
impacts; they are increasingly being recognized as a critical feature of every stormwater BMP
plan, especially with regard to site design and development and maintenance costs. The key
benefit of some nonstructural practices is that they can also reduce the generation of stormwater
runoff from the site, thereby reducing the size and cost of storage; in addition to providing partial
removal of many pollutants (MDE, 2000). A comprehensive plan addressing urban stormwater
2-4
runoff pollution prevention and control requires the implementation of a combination of
nonstructural and structural BMPs for existing and new development.
Nonstructural or source control approaches for stormwater management comprise three major
components:
<
planning, design and construction of developments and re-developments to minimize or
eliminate adverse impacts;
<
good maintenance of impervious and pervious surfaces to minimize exposure and release
of pollutants; and
<
education and training to promote awareness of the potential problems associated with
urban stormwater runoff and of specific BMPs to help solve problems (WEF and ASCE,
1998).
Typically, nonstructural BMPs for urban areas can be grouped into categories as shown in Table
2-1 (WEF and ASCE, 1998).
Table 2-1. Nonstructural BMPs for Urban Stormwater Runoff
Major Categories
Nonstructural Practice
2.1.1 Public Education
<
Public Education and Outreach
2.1.2 Planning and
Management
<
<
<
<
<
Better Site Design
Vegetation Controls
Reduction/Disconnection of Impervious Areas
Green Roofs*
Low-Impact Development**
2.1.3 Materials Management
<
<
Alternative Product Substitution
Housekeeping Practices
2.1.4 Street/Storm Drain
Maintenance
<
<
<
<
<
<
Street Cleaning
Catchbasin Cleaning
Storm Drain Flushing
Road and Bridge Maintenance
BMP Maintenance
Storm Channel and Creek Maintenance
2.1.5 Spill Prevention and
Cleanup
<
<
Above Ground Tank Spill Control
Vehicle Spill Control
2.1.6 Illegal Dumping
Controls
<
<
<
<
Illegal Dumping Controls
Storm Drain Stenciling
Household Hazardous Waste Collection
Used Oil Recycling
2-5
2.1.7 Illicit Connection
Control
<
<
<
Illicit Connection Prevention
Illicit Connection - Detection and Removal
Leaking Sanitary Sewer and Septic Tank Control
2.1.8 Stormwater Reuse
<
<
<
<
Landscape Watering
Toilet Flushing
Cooling Water
Aesthetic and Recreational Ponds
* Also considered a structural BMP since engineering principles and design are involved
**C on sid ere d a c ombin atio n o f b oth n on stru ctura l a nd structura l B M P
(WEF and ASCE, 1998)
2.1.1 Public Education
Public understanding, involvement and support is an essential component of stormwater
management programs. A public education and participation plan provides the municipality with
a strategy for involving the public in making stormwater management decisions, and for
educating its employees, the public and businesses about the importance of protecting stormwater
from improper use, storage and disposal of pollutants. The intent of programs targeting public
involvement is to get the public behind the overall stormwater management program to increase
its acceptance. These programs are designed to raise public awareness on diffuse source
pollution and encourage people to change their everyday behavior in the use and disposal of
chemicals, other household and automotive products in order to minimize and/or prevent the
entry of these products into stormwater runoff. Essentially, such programs include conducting
workshops, open houses, surveys and other means of including the public in providing input into
and commenting on stormwater management efforts. This may be accomplished by distributing
educational materials to the community or by conducting equivalent outreach activities about the
impacts of stormwater discharges on water bodies and the steps the public can take to reduce
pollutants in urban stormwater runoff (U.S. EPA, 2001). This promises to be a cost-effective
way of stormwater quality management; yet, its effectiveness on the actual reductions of target
constituents in receiving waters has yet to be definitively demonstrated (WEF and ASCE, 1998).
The objectives of a public education and participation plan should be to:
<
promote clear identification and understanding of the problem and the solutions,
<
identify responsible parties and efforts to date,
<
promote community ownership of the problem and its solutions,
<
change behaviors, and
<
integrate public feedback into program implementation.
The main components of a public education program include the following (U.S. EPA, 1999b).
<
Public outreach/education for homeowners
•
Lawn and garden activities education
•
Water conservation practices for homeowners
2-6
•
•
•
<
<
<
Pet waste management education
Proper disposal of household hazardous wastes
Trash management education
Targeting public outreach/education
Public outreach programs for new development
Pollution prevention programs for existing development
2.1.2 Planning and Management of Developing Areas
These practices by local governments and large land owners/developers are aimed at reducing
urban stormwater runoff and the discharge of pollutants through stormwater from new
developments, and are most effective when applied during the site-planning phase of new
development. This BMP presents an important opportunity to reduce the pollutants in
stormwater runoff by using a comprehensive planning process to control or prevent certain land
use activities in areas where water quality is sensitive to development. It is applicable to all types
of land use and represents one of the most effective pollution prevention practices. Land use
planning and management are critical to watershed management.
2.1.2.1 Better Site Design
The use of better site design techniques is one of the few watershed management practices that
seeks to simultaneously reduce pollutant loads, conserve natural areas, save money and increase
property values, and at the same time it collectively employs a variety of methods to accomplish
three goals at every development site, to: (i) reduce the amount of impervious cover and/or
directly connect impervious cover; (ii) increase natural lands set aside for conservation; and, (iii)
use pervious areas for more effective stormwater treatment (CWP, 2000). However, in many
communities, it is observed that many of the regulations governing the development process in
such areas as zoning, parking and street standards, and drainage, to name a few, are at crosspurposes with better site design and need to be addressed. Better site design practices identify
areas where existing codes and standards can be changed to better protect streams, lakes, and
wetlands at the local level, and fall into the following three categories.
<
narrow residential streets and smaller parking lots such as minimizing residential street
cul-de-sac sizes, reducing overall imperviousness associated with parking lots;
<
lot development such as open space design development, routing of rooftop runoff to
pervious areas; and
<
conservation of natural areas such as buffer zones, setbacks and easements.
2.1.2.2 Vegetation Control
This typically involves a combination of mechanical methods and careful application of
chemicals (fertilizers, insecticides and herbicides). Mechanical vegetation controls include
practices such as leaving existing vegetation, less frequent cutting, hand cutting, planting
low-maintenance vegetation, collecting and properly disposing of clippings and cuttings, and
2-7
educating the public and employees. Stormwater quality strongly impacts areas including steep
slopes, vegetated drainage channels, creeks, areas adjacent to catch basins, detention/retention
basins. Flat or relatively flat vegetated areas, areas not adjacent to drainage structures, and areas
screened from drainage structures or from vegetation are less impacted (WEF and ASCE, 1998).
2.1.2.3 Reduction/Disconnection of Impervious Areas
The volume of stormwater runoff generated in a development can be greatly reduced by
minimizing the amount of impervious surfaces. Reductions in impervious area can be
undertaken by reducing the overall size of the developed area, and/or by reducing the amount of
impervious surface created within the developed area. Disconnection of impervious surfaces can
be undertaken by directing runoff from roofs and paved surfaces over vegetated surfaces before it
reaches the drainage conveyance system (GVSDD, 1999). Reductions in impervious area can
also be achieved through cluster developments that maximize open (undeveloped) space and
minimize the required length of roadway and other infrastructure. Clustering concentrates
development on smaller lots into compact areas and leaves relatively large areas undeveloped, in
contrast to conventional grid developments that cover the entire site with larger lots and result in
more overall impervious area (roads, sidewalks etc). This approach will help address peak flow
control, stream bank erosion protection, removal of drainage path obstruction, water quality
enhancement, groundwater recharge and community enhancement. Clustering is suitable for new
developments and redevelopment areas, but has limited opportunities for use in areas of existing
development. A few measures that reduce impervious surfaces include:
<
reducing the width of streets,
<
limiting the length and radius of cul-de-sacs,
<
using porous pavement of modular block pavers in parking areas and low-traffic areas,
<
placing sidewalks to only one side of the street or no sidewalks where possible, and
<
reducing frontage requirements to lessen paved surface areas (U.S. EPA, 1999b).
2.1.2.4 Green Roofs
Green roofs, also known as vegetated roof covers, eco-roofs or nature roofs, are multi-beneficial
structural components that help to mitigate the effects of urbanization on water quality by
filtering, absorbing or detaining rainfall (GreenRoofs; Liptan and Strecker, 2003). In areas of
high-density development, where pervious surfaces and open ground often make up 10% or less
of total surface area capable of absorbing or diverting storm water runoff, green roofs are one of
the best ways to reduce runoff volumes via evapotranspiration losses. In addition, they can
provide significant social, environmental and financial benefits. There has been a lot of
discussion with due explanation, about the need to describe green rooftops as a structural or a
nonstructural BMP; this chapter examines green roofs as a nonstructural BMP, mainly for the
reason that the goal of green rooftops is runoff volume reduction, pollution prevention and
mitigation at the source (hydrological source control) with minimal and cost-effective landscape
features located at the lot level, as opposed to structural controls that convey, manage and treat
stormwater in large, expensive, end-of-pipe facilities.
2-8
Green roof systems involve the creation of a "contained" green space atop a human-made roofing
system that include the roof structure and insulation; a waterproofing membrane, often with root
repellent; a drainage layer; landscape or filter cloth to contain roots and soil; a specialized
growing medium (which may not include soil); and plants (MWLAP, 1992). They comprise an
impermeable membrane or similar structure that supports a lightweight soil medium and living
vegetation, e.g., grass or groundcover placed on all or part of the building roof (Metro Council,
2001). The soil and vegetation layer provide a means of replacing the impermeable surfaces of
building roofs to reduce stormwater runoff volumes, control stormwater peak flows, improve
stormwater quality, and reduce stormwater runoff temperature. The soil is planted with a
specialized mix of plants that can thrive in the harsh, dry, high temperature conditions of the roof
and tolerate short periods of inundation from storm events. Care should be exercised about the
soils employed to be sure that they do not become a source of pollutants, for example, nutrients.
Historically, engineered green roofs originated in northern Europe, where sod roofs and walls
have been utilized as construction materials for hundreds of years. The development of
contemporary approaches to green roof technology began in the urban areas of Germany over 30
years ago. Despite centuries of use in Iceland and various research initiatives to encourage their
use in Canada, Germany, France, Austria and Switzerland, green roof tops are only a recent entry
in the U.S. However, several new and planned projects, including a retrofit of the Chicago City
Hall, a 450,000 ft2 vegetated roof planned for a new Ford assembly plant in Dearborn, MI, a
vegetated roof at the Green Institute in Minneapolis, a vegetated roof top at the Fencing
Academy, Philadelphia (U.S. EPA, 2000c), numerous green roofs in Portland, OR, encouraged
by a city incentive program (Liptan and Strecker, 2003), and the Green Roof research center at
the University of Pennsylvania dedicated to demonstrating and promoting green roof research,
education and technology transfer, are raising visibility and encouraging the nascent U.S. market
for green roof technology.
Green roofs are especially effective in controlling intense, short-duration summer storms and
have been shown to reduce cumulative annual runoff by 50% in temperate climates with summer
storm reductions of up to 80 or 90%. Green roof systems partially restore the natural hydrologic
cycle, compensating for the impervious surface areas associated with urban development. These
systems absorb and retain precipitation; the water is used by the plants or slowly released to the
storm system. In general, green roof systems provide summer retention rates of 70 to 100% and
winter rates of 40 to 50% depending on factors such as precipitation patterns and forms,
substrate, type and depth of vegetation, temperature, humidity, sun and wind. In addition, they
moderate the temperature of runoff and filter nutrients, sediments and metals from the runoff.
Green roofs provide stormwater management benefits by:
<
utilizing the biological, physical, and chemical processes found in the plant and soil
complex to prevent airborne and rain-entrained pollutants from entering the storm drain
system; and
<
reducing the runoff volume and peak discharge rate by holding back and slowing down
the water that would otherwise flow quickly into the storm drain system.
2-9
The hydrologic processes that can be influenced by design choices and aid in managing
stormwater include: interception of rainfall by foliage and subsequent evaporation; soil moisture
penetration, soil adsorbing ability, and soil moisture maintenance (e.g., irrigation); reduction in
the velocity of runoff; shallow subterranean flow through the soil; root zone moisture uptake;
and, evapotranspiration (U.S. EPA, 2000a).
2.1.2.5 Low-Impact Development (LID)
LID is a site design strategy with the goal of maintaining or replicating or minimizing the change
in the pre-development hydrologic regime through the use of design techniques to create
functionally equivalent hydrologic landscapes. Hydrologic functions of storage, infiltration, and
ground water recharge, as well as the volume and frequency of discharges are maintained, or
their changes reduced through the use of integrated and distributed micro-scale stormwater
retention and detention areas, reduction of impervious surfaces, and the lengthening of flow paths
and runoff time (U.S. EPA, 2000d). Over the last several years, a number of techniques when
applied in combination in a methodological way, have asserted to result in little or no stormwater
impacts. These techniques began to be increasingly implemented in stormwater management
efforts in the early 1990's and were given the term Low-Impact Development (LID) by Prince
George’s County (Strecker, 2001).
The principles of LID are based on controlling stormwater at the source by the use of micro scale
controls that are distributed throughout the site. These are multi functional site design practices
and incorporate alternative stormwater management practices such as functional landscapes that
act as stormwater control facilities, depression storage and open drainage swales. Such systems
of controls can reduce or eliminate the need for a centralized BMP facility for the control of
stormwater runoff. LID measures provide a means to address both pollutant removal and the
protection of pre-development hydrologic functions. Some basic LID principles include
conservation of natural features, minimization of impervious surfaces, hydraulic disconnects,
disbursement of runoff, and phytoremediation. LID practices such as bioretention facilities or
rain gardens, grass swales and channels, vegetated rooftops, rain barrels, cisterns, vegetated filter
strips and permeable pavements perform both runoff volume reduction and pollutant filtering
functions (Prince George's County (DER), 2002). The benefits of LID are many:
<
LID addresses hydrologic changes caused by development at the site level, which reduces
the downstream impact of increased imperviousness.
<
LID practices, when used in combination with each other and with traditional treatment
practices such as regional retention ponds, further reduce pollutant loading to receiving
water bodies.
<
Many LID practices involve natural landscaping, including the planting of trees, shrubs,
and flower gardens that enhance site aesthetics and reduce mowing requirements; the
choice of local species and locally grown stocks can greatly reduce watering and fertilizer
requirements.
<
Careful regrading and well-sited localized depression and storage areas with slower but
assured drainage help improve overall site drainage, prevent overly long pooling and
2 - 10
creation of mosquito-breeding habitat, and reduce both onsite and downstream flooding.
2.1.3 Materials Management
The responsible management of common chemicals, such as fertilizers, solvents, paints, cleaners,
and automotive products can significantly reduce polluted stormwater runoff (U.S. EPA, 2001).
Such products must be handled properly in all stages of their useful lives. These practices
include controlling the use, storage and disposal of chemicals that could pollute stormwater
runoff with the objective of reducing the opportunity for rainfall or runoff to come into contact
with these chemicals. This BMP includes activities pertaining to material use controls, material
exposure controls, and material disposal and recycling controls. Material storage controls help
prevent or reduce the discharge of pollutants to stormwater from material delivery and storage
areas for municipal and commercial operations. This BMP primarily concerns the design and
maintenance of material storage areas that reduce exposure to stormwater by storing materials
inside or under cover on paved surfaces; using secondary containment, where required;
minimizing storage and handling of hazardous materials; and, regularly inspecting storage areas
(WEF and ASCE, 1998). Materials management primarily includes use minimization and the
use of less toxic alternatives and recommends the following approaches as control measures:
<
using products only when needed,
<
using chemicals according to label instructions or lesser quantities, and
<
applying only when required.
2.1.3.1 Alternative Product Substitution
The presence of toxic substances in stormwater and receiving waters can be drastically reduced
by prudent use of alternatives to toxic substances. The most common toxic substances found in
homes and small businesses include cleaners, automotive products, pesticides, fertilizers, paints
and fuels that find their way into ground and surface receiving waters upon improper disposal.
Alternatives exist for most product classes; some examples include rechargeable batteries, baking
soda, pump-type or non-aerosols etc. Most garden products have natural alternatives, and less
toxic alternatives exist for home and automotive repair products (WA DOE, 2001). For example,
alternatives to pesticides, such as insecticidal soap or natural bacteria recommended in integrated
pest management programs can help reduce the need for pesticides (U.S. EPA, 1999b).
Integrated pest management strategies include the use of natural predators and pathogens,
mechanical controls, native and pest-resistant plants, removing pest habitat and localized use of
appropriate use of chemicals as a last resort when problems are observed (Taylor and Wong,
2002).
Care should also be exercised in selecting building materials for new construction and
renovation. Commonly used building materials such as zinc roof gutters and architectural copper
can leach significant quantities of dissolved metals into runoff causing adverse ecological
impacts. The use of proper alternative materials can reduce potential contaminants in runoff by
eliminating compounds that leach into runoff. This will reduce the need for pesticide
2 - 11
application, painting and other maintenance by reducing the volume of runoff. Alternative
building materials are available instead of treated copper base (or other treatment of wood) as
lumber for decking (plastic or plastic-wood composites), roofing materials (coated metal
products, roofing materials made of recycled rubber or plastic instead of metals), home siding
(vinyl, cement-fiber), and paving for driveways, decks and sidewalks (porous concrete and
asphalt, modular blocks and crushed granite). Other examples include the use of formaldehydefree plywood, and low VOC paints, carpets, and pads (CASQA, 2003).
The promotion of safer alternative products, however, should be coupled with other programs
designed to reduce the presence of hazardous or toxic materials in homes, offices, commercial
spaces and more importantly, urban stormwater runoff. Examples of such programs are
hazardous materials collection, good housekeeping or material management practices, oil and
automotive waste recycling, and spill response and prevention (WEF and ASCE, 1998).
Examples of proactive communities (those that have instituted effective programs promoting
safer alternative products) include Santa Clara County, the city of Palo Alto, the city and county
of San Francisco, CA, and the cities of Bellevue and Seattle, WA.
2.1.3.2 Housekeeping Practices
The promotion of efficient and safe housekeeping practices (which include storage, use, cleanup,
and disposal) when handling potentially harmful materials such as pesticides, fertilizers, cleaning
solutions, paint products, automotive products, and swimming pool chemicals can be an effective
source control BMP. The target audiences for making this BMP an effective one are municipal
employees, the general public and small and large businesses. This BMP should be implemented
in conjunction with safer alternative products BMPs and integrated to the highest extent possible
with existing programs in the municipality.
Commercial and retail areas can contribute significantly to pollutant loadings in urban
stormwater runoff, with the largest contribution from impervious surfaces used for vehicle
parking, storage and maintenance areas that contribute metals, sediments and hydrocarbons.
Good housekeeping practices include using porous pavement or modular paving systems for
vehicle parking lots, minimizing or eliminating exposure of materials and equipment to rainfall
by storing inside or under cover, using dry cleanup techniques instead of wet techniques,
minimizing pesticide/herbicide and fertilizer use, limiting discharges of equipment wash water to
storm drains, and limiting direct runoff of rooftops to storm drains (U.S. EPA, 1993a, 1999b).
2.1.4 Street/Storm Drain Maintenance
This BMP category applies to the removal of pollutants from paved areas and the maintenance of
runoff quality controls that exist within the drainage system. Examples include street and
parking lot sweeping, catch basin cleaning, road and bridge maintenance, and maintenance of
structural controls in the system for runoff quality management. This group also includes the use
of good housekeeping measures whenever performing pavement maintenance such as installing
2 - 12
asphalt overlays or conducting seal and chip procedures.
2.1.4.1 Street Cleaning
Street and parking area cleaning was extensively studied during early U.S. EPA-funded research
projects since the 1970s, as it was thought to be an effective means of controlling the quality of
runoff associated with large quantities of pollutants found on streets and parking lots (Field and
Sullivan, 2003). This management measure involves employing pavement cleaning practices
such as street sweeping on a regular basis to minimize pollutant export to receiving waters.
These cleaning practices are designed to remove from road and parking lot surfaces sediment
debris and other pollutants that are potential sources of pollution affecting urban waterways
(FHWA, 2000). Street sweeping is considered to be an effective ultra-urban best management
practice for reducing total suspended solids and associated pollutant washoff from urban streets.
This is especially well suited to those urban environments with little land available for the
installation of structural controls. Areas of application include commercial business districts,
industrial sites, and intensely developed areas in close proximity to receiving waters. In highway
applications, street-sweeping may be considered for road shoulders (where safety permits), rest
stop parking areas, or maintenance yards. There are many kinds of street sweepers currently
being used: mechanical; vacuum-assisted sweepers; tandem sweeping; regenerative air sweepers;
and, vacuum-assisted dry sweepers.
Performance evaluation of street sweeping as a stormwater quality BMP under the NURP
Program (U.S. EPA, 1983) concluded that while this technique was effective in removing litter
and coarse fractions of sediment from the surface, it provided no significant reduction in nutrient
concentrations in stormwater runoff. However, newer street sweeping technology more
effectively removes the finer fraction of suspended particles that carry a substantial portion of the
stormwater pollutant load; sweeping programs carried out routinely on a weekly or bimonthly
basis have shown potential reductions of up to 80% in annual loads of total suspended solids and
associated pollutants (FHWA, 2000; U.S. EPA, 2001).
Although street sweeping has been widely practiced for litter and dust control, its implementation
as a stormwater pollution control practice is a fairly recent development (NVPDC, 1996). “The
effectiveness of street sweeping programs depends more on factors such as land-use activities,
the inter-event dry period, street sweeping frequency and timing, access to source areas and
sweep operation than the actual street sweeping mechanism” (Walker and Wong, 1999, in Taylor
and Wong, 2002). The performance of street sweepers is significantly affected by factors that
include street dirt loadings, street texture, litter and moisture, parked car conditions, and
equipment operating conditions. It must be noted that extensive street cleaning might be more
beneficial in reducing SS and heavy metals in arid western regions of the U.S., where the
infrequent rains allow substantial pollutant accumulation on the street; however, in the wet
southeastern U.S. where large and frequent rains occur, “street cleaning is likely to have much
less direct water quality benefit, beyond possibly important litter and floatable control” (Field
and Sullivan, 2003).
2 - 13
2.1.4.2 Catchbasin Cleaning
Catchbasins are chambers or sumps, usually built at the curb line, that allow surface water runoff
to enter the stormwater conveyance system. Many catchbasins have a low area below the invert
of the outlet pipe intended to retain coarse sediment. Catchbasins as BMPs are reasonably
effective in protecting sewers from receiving loads of very coarse solids greater than 0.04 in.
diameter, but not very effective in capturing fine particulates such as clays or silt (MDEQ, 1998).
Catchbasins naturally accumulate sediment and debris such as trash and leaf litter; sediment
trapping by catch basins prevents solids from clogging the storm sewers and being washed into
receiving waters. Uncleaned catchbasins which allow coarse materials to overflow may
contribute higher loads of BOD (biochemical oxygen demand) and sediment to the receiving
streams. Catchbasins in order to be effective require periodic cleaning using either a vacuum or
adductor to remove the accumulated pollutants and maintain the pollutant removal efficiency
(U.S. EPA, 1999a, 1999b). One recommendation is to clean the catchbasins at least two times
per yr (just before and after the rainy season), and/or when the catchbasin storage is one-third
full, whichever happens first, as this will help keep pollutants and sediments from re-entering
stormwater (Pierce County). Some limitations associated with catchbasin cleaning include:
<
Catchbasin debris usually contain appreciable amounts of water and “offensive”
(odorous) organic matter that must be properly disposed of.
<
Catchbasins may be difficult to clean in areas with poor accessibility, traffic congestion
and parking problems.
<
Cleaning is difficult in winter in the presence of snow and ice (U.S. EPA, 1999a).
Sediment and debris removed from catchbasins can potentially be classified as hazardous waste
particularly when cleaning is infrequent and/or in industrial areas and always needs to be
disposed of properly to avoid negative environmental impacts.
2.1.4.3 Roadway and Bridge Maintenance
Road and street surfaces undergo breakdown due to frictional action of traffic, freeze-thaw
temperatures, frost heaving, ultraviolet degradation and erosion of road subbase. This results in
exposure of unstabilized subbase material to erosive forces of water and subsequent increases in
suspended solids concentration as well as other constituents such as PAHs. The substantial
loadings of sediments and other pollutants generated during daily roadway and bridge use and
scheduled repair operations pose a threat to local water quality by contributing heavy metals,
hydrocarbons, PAHs, sediment and debris to stormwater runoff (U.S. EPA, 2001). A few
measures that help to alleviate the impact of pollutants from roadway surfaces include:
<
Routine performance of maintenance activities such as sweeping, vegetation
maintenance, and cleaning of runoff control structures.
<
Modifications in roadway resurfacing practices and application techniques for salt and
other deicers.
Extensive studies on the characterization and environmental impacts of highway deicing
2 - 14
chemicals have been conducted by the U.S. EPA and the negative environmental impacts of these
salting and sanding operations is well documented (D'Itri, 1992). Alternative de-icing products
such as acetates, formates and agricultural residues can be used if impacts due to traditional
deicing products are significant (U.S. EPA, 1999a, 1999b)
2.1.4.4 Storm Drain Flushing
Overflowing of storm drains and storm drain inlets, as well as the resulting increase in erosion
are natural consequences of irregular cleaning and maintenance of these facilities. Routine
cleaning of storm drains is associated with a lot of benefits, including increased dissolved
oxygen, reduced levels of bacteria and support of instream habitat. Flushing a storm drain with
water to suspend and remove deposited materials is particularly beneficial for storm drain pipes
with relatively flat grades or low flows where self-cleansing becomes difficult (U.S. EPA, 2001).
Flushing helps to ensure that pipes convey design flow and removes pollutants from the storm
drain system (WEF and ASCE, 1998).
2.1.4.5 BMP Maintenance
BMPs require a variety of periodic maintenance activities in order to enhance performance. This
includes sediment removal, vegetation maintenance, periodic maintenance and repair of outlet
structures if needed, periodic replacement of filter media, to name a few (U.S. EPA, 1999b).
Regular inspection of control measures is essential in order to maintain the effectiveness of
post-construction stormwater BMPs. The inspection and maintenance of BMPs can be
categorized into two groups: expected routine maintenance, and non-routine (repair)
maintenance. Routine maintenance involves checks performed on a regular basis to keep the
BMP in good working order and aesthetically pleasing and is an efficient way to avoid the health
and safety threat inherent in BMP neglect (e.g., prevent potential nuisance situations, reduce the
need for repair maintenance, reduce the chance of polluting stormwater runoff by finding and
correcting problems before the next rain) (U.S. EPA, 2001). A general guideline on BMP
maintenance is presented in more detail in Chapter 5 of this white paper.
2.1.4.6 Storm Channel and Creek Maintenance
Regular removal of illegally dumped items and material from storm drainage channels and creeks
helps to reduce pollutant levels in stormwater runoff. The approaches for effective storm channel
and creek maintenance include the following:
<
identify and regularly clean up stormwater "hotspots" and other storm drainage areas
where illegal dumping and disposal occurs,
<
establish and maintain buffer zones along creeks, and
<
modify storm channel characteristics: e.g., improve channel hydraulics; reduce channel
erosion and increase pollutant removals; and enhance aesthetics and habitat values.
2 - 15
2.1.4.7 Stormwater “Hotspots”
Stormwater “hotspots” are areas of the urban landscape that often generate higher concentrations
and/or loads of certain pollutants, such as hydrocarbons, trace metals or toxicants than are
normally found in urban stormwater runoff and that are termed so based on monitoring studies
(MDE, 2000; Atlanta Regional Commission, 2001). These areas merit special management and
the use of specific pollution prevention activities and/or structural stormwater controls. These
are areas where land use or activities generate potentially highly contaminated stormwater runoff,
with concentrations of pollutants in excess of those typically found in urban stormwater.
The designation of a site as a stormwater hotspot has important implications for how the
stormwater is managed, the most important one being that untreated stormwater runoff from
hotspots cannot be allowed to infiltrate into groundwater where it may contaminate water
supplies. Secondly, a higher level of stormwater treatment is needed at hotspot sites to prevent
pollutant washoff after construction. This typically involves preparing and implementing a
stormwater pollution prevention plan that involves a series of operational practices at the site to
reduce the generation of pollutants by preventing contact with rainfall. The following land uses
and activities are typically considered as stormwater hotspots when exposed to stormwater (VA
DCR, 1999; Atlanta Regional Commission, 2001; CASQA, 2003):
<
vehicle salvage yards and recycling facilities;
<
vehicle fueling, service, and maintenance facilities;
<
vehicle and equipment cleaning facilities;
<
fleet storage areas;
<
industrial areas;
<
marinas - service and maintenance;
<
outdoor liquid container storage;
<
outdoor loading/unloading facilities;
<
public works storage areas;
<
facilities that generate or store hazardous materials;
<
commercial container nurseries; and
<
other land uses and activities as designated by an appropriate review authority.
2.1.5 Spill Prevention and Cleanup
This category of BMPs includes programs that reduce the risk of spills during outdoor handling
and transportation of chemicals and other materials. It also includes the development of plans
and programs to respond, contain and rapidly clean up spills when they do occur so that they do
not enter the storm drain system. According to the U.S. EPA (2001), spill response and
prevention plans should “clearly state measures to stop the source of a spill, contain the spill,
cleanup the spill, dispose of contaminated materials, and train personnel to prevent and control
future spills.” Such plans are most applicable to construction sites where hazardous wastes are
stored or used (U.S. EPA, 1992). The preliminary steps include: (i) identifying potential spill or
source areas such as loading and unloading, storage, and processing areas; places that generate
2 - 16
dust or particulates; and, areas designated for waste disposal; and, (ii) evaluating stationary
facilities that include manufacturing areas, warehouses, service stations, parking lots, and access
roads.
2.1.5.1 Above Ground Tank Spill Control
Preventing or reducing the discharge of pollutants to stormwater from storage tanks can be done
by installing safeguards against accidental releases, installing secondary containment, conducting
regular inspections, and training employees in standard operating procedures and spill cleanup
techniques (WEF and ASCE, 1998).
2.1.5.2 Vehicle Spill Control
Methods for preventing or reducing the discharge of pollutants to stormwater from vehicle leaks
and spills include reducing the chance for spills by preventive maintenance, stopping the source
of spills, containing and cleaning spills, properly disposing of all spill materials, and training
employees. The following practices help to reduce the impact from these activities (U.S. EPA,
1999b):
<
All spills or leaks should be cleaned using a dry absorbent such as cat litter or
commercially available sorbents and disposed of appropriately.
<
All used fluids should be recycled or disposed of appropriately.
<
All fluid leaks should be repaired as soon as possible to reduce discharge to the
environment.
2.1.6 Illegal Dumping Controls
The use of measures to detect, correct, and enforce laws against illegal dumping of pollutants in
gutters and streets and into the storm drain system and creeks can have a significant effect on
stormwater quality. This BMP includes the control of both direct and indirect sources (WEF and
ASCE, 1998). This category of BMPs comprises ordinances, public education programs, and
authorized enforcement measures aimed at keeping individuals and businesses from dumping
various waste products onto the urban landscape and into the drainage system. The control of
illegal dumping practices is important to prevent contaminated runoff from entering wells and
surface water, as well as averting flooding due to blockages of drainage channels for runoff (U.S.
EPA, 2001). The key to successfully using this BMP is increasing public awareness of the
problem and its implications. Some of the issues that need to be examined when creating a
program include:
<
the locations of persistent illegal dumping activity;
<
types of wastes dumped and the profile of dumpers;
<
possible driving forces behind illegal dumping, such as excessive user fees, restrictive
curbside trash pickup, or ineffective recycling programs;
<
previous education and cleanup efforts;
<
current control programs and local laws or ordinances addressing the problem; and
2 - 17
<
sources of funding and additional resources that may be required.
Illegal dumping controls should focus on the following program areas (U.S. EPA, 1998a): (i)
cleanup efforts; (ii) community outreach and involvement; (iii) targeted enforcement; and, (iv)
tracking and evaluation. A few well known examples of illegal dumping controls are discussed
below:
2.1.6.1 Storm Drain Stenciling
Storm drains frequently discharge runoff directly to water bodies with little or no treatment.
Dumping of waste materials into these systems (inlets, catchbasins, channels and creeks) could
have severe impacts on receiving water quality. Storm drain stenciling programs that educate
residents not to dump materials into storm drains or onto sidewalks, streets, parking lots and
gutters is an effective means of reducing nonpoint source pollution associated with such illegal
dumping. Storm drain signs and stencils use prohibitive language and/or graphic icons
discouraging illegal dumping of improper materials into the urban stormwater runoff conveyance
system, and are typically placed directly adjacent to storm drain inlets (WEF and ASCE, 1998;
U.S. EPA, 1999b; CASQA, 2003). The use of these highly visible source controls is a common
educational strategy in pollution prevention in the U.S. and Australia. Although a few studies
reported a positive correlation between seeing the stencils and levels of stormwater awareness,
there is no conclusive observation to demonstrate that stormwater drain stenciling raises public
awareness or induces behavioral change (Taylor and Wong, 2002).
2.1.6.2 Household Hazardous Waste Collection
Collection programs to recover substances that may otherwise be dumped onto land or into
stormwater is a common BMP in major areas. The U.S. EPA (1993a) reported rapid growth in
the number of household hazardous chemical collection programs in the U.S., from a mere 2
programs in 1980 to 822 in 1990. Collection efforts should be integrated with an already
established municipal solid waste program. The frequency of collection is based on waste type,
community characteristics, existing programs and budgets.
2.1.6.3 Used Oil Recycling
Used motor oil is a hazardous waste as it contains heavy metals picked up from the engine during
use and should be disposed of at a local recycling or disposal facility. The recycling of used
motor oil is a responsible alternative to improper disposal practices such as dumping oil in the
sanitary sewer or storm drain system, applying oil to roads for dust control, placing used oil and
filters in the trash for landfill disposal, or simply pouring used oil on the ground. Used oil can be
recycled in a number of ways (U.S. EPA, 2001):
<
Reprocessing used oil into fuel, which is used for heating and cooling homes. This is the
most common method of recycling used oil in the U.S. with approximately 750 million
gallons being reprocessed every year and marketed to asphalt plants, steel mills, boilers,
2 - 18
<
<
<
<
<
pulp and paper mills, cement/lime kilns, and a number of other places.
Motor oil can also be burned in furnaces for heat or in power plants to generate electricity
for homes, businesses, or schools.
It can also be blended for marine fuels, mixed with asphalts for paving, or used in
industrial burners.
Used motor oil can also be used in specially designed municipal garages, space heaters,
and automotive bays.
Used motor oil can be re-refined into lubricating oils that meet the same standards as
virgin/new oil.
Implementation of recycling programs involving public participation at local and national
levels
2.1.7 Illicit Connection Controls
Illicit connections are defined as "illegal and/or improper connections to storm drainage systems
and receiving waters" (CWP, 1998). The interest in illicit or inappropriate connections to storm
drainage systems is an outgrowth of investigations into the larger problem of determining the role
of urban stormwater runoff as a contributor to receiving water quality problems. Identifying
illicit and improper connections is necessary for all sewer systems, especially in areas where
pollutants with unknown sources have been detected in receiving waters. Non-stormwater outfall
discharges fall into three categories: (i) pathogenic/toxicant; (ii) nuisance and aquatic life
threatening; and, (iii) clean water. Outfall discharges containing pathogenic or toxic pollutants
are the most important, and the most likely sources of these pollutants are sanitary and industrial
wastewaters (U.S. EPA, 1993c). Potential sources of contaminated entries into storm drainage
systems include residential and commercial sources; industrial sources; intermittent sources;
direct connections to storm drains; and infiltration to storm drains. A major source of illicit
discharges to storm drain systems are direct connections of sanitary sewer piping to the storm
drain system. Besides direct connections, seepage and sewage from leaking sanitary sewer lines
find their way into storm drains, especially in areas where storm drains run parallel to the sanitary
sewer lines. The level and types of industrial activities and the surrounding land uses and
ordinances will affect the methods used to identify illicit connections.
2.1.7.1 Illicit Connection - Prevention, Detection and Removal
This group of controls is directed at preventing by ordinance, and eliminating by discovery or
removal, connections to the storm drainage system that discharge any material other than
stormwater runoff. Bans on connection of floor drains, wash down areas, septic tank overflows,
and other similar practices to the stormwater conveyance system are all a part of this BMP
category. The prevention of unwarranted physical connections to the storm drain system from
sanitary sewers and floor drains through regulation, regular inspection, testing and education can
remove a significant source of stormwater pollution. Control procedures for detecting and
removing illegal connections from the storm drain conveyance system should be implemented to
identify, repair, and remediate infiltration, inflow, and wet-weather overflows from a sanitary
2 - 19
sewer to the storm drain conveyance system. This will effectively reduce or prevent
unauthorized discharges to receiving waters; some strategies include field-screening, follow-up
testing and complaint investigation (U.S. EPA, 1993c, 2003).
Several methods exist for the detection and elimination of illicit cross-connections, which
significantly reduce the concentrations of bacteria, nutrients, and oxygen demanding substances
contained in stormwater discharges. Useful indicators of the presence of cross connections
include dry-weather flows in storm sewer lines and biological indicators that indicate the
presence of human fecal matter in storm drain outfalls (U.S. EPA, 1999b). Excavation and
correction of illicit connections must be a natural follow-up to detecting these in the first place,
and additionally, plans for new development need to be reviewed with periodic inspections
during construction preventing future cross connections from being placed.
2.1.7.2 Failing Septic Systems and Sanitary Sewer Overflows
A failing septic system is considered to be one that discharges effluent with pollutant
concentrations exceeding established water quality standards. Failure rates for septic systems
typically range between 1 and 5 % each year but can be much higher in some regions (U.S. EPA,
2001). Septic system failure has several causes: (i) unsuitable soil conditions; (ii) improper
design and installation; or (iii) inadequate maintenance practices. Improperly functioning septic
systems contribute significant loads of pollutants (especially nitrogen) and microbial pathogens.
Identifying and eliminating these control the contamination of ground and surface water supplies
from untreated wastewater discharges.
Sanitary sewer overflows (SSOs) occur in urbanized areas where a separate sanitary sewer
system has been created to move wastewater from households and businesses to treatment plants.
The detection and elimination of SSOs is most important because sanitary sewer collection
systems represent a significant investment for urban municipalities. SSOs can often be reduced
or eliminated by a number of practices, including the following (U.S. EPA, 2001):
<
sewer system cleaning and maintenance,
<
reducing infiltration and inflow through rehabilitation and repair of broken or leaking
sewer lines,
<
enlarging or upgrading the capacity of sewer lines, pump stations, or sewage treatment
plants,
<
constructing wet-weather storage and treatment facilities to treat excess flows, and
<
addressing SSOs during sewer system master planning and facilities planning.
2.1.8 Stormwater Reuse
Stormwater runoff stored in a surface pond or in the surficial aquifer and then used as a source of
irrigation water can reduce potable water use in an area and reduce pollutant loadings from
stormwater. The reuse of stormwater is an alternative source of non-potable water and most
2 - 20
likely the major use is for irrigation. Stormwater reuse can take place at the (i) household; (ii)
municipal; or, (iii) a larger (regional)level (UNEP IETC, 2000).
At the household level, roof runoff could be collected in a tank for use as drinking water
(common in arid regions), flushing toilets, or irrigation of gardens. The first flush, generally
contaminated with dust particles, leaf litter and animal droppings, can be diverted using a simple
diverter; gross particles should be filtered by placing a screen near the inlet. Directing water
from the roof directly to garden bed rather than through soakways would benefit shallow-rooted
vegetation, especially in arid regions.
At the municipal level, stormwater can be stored in ponds for use for irrigation of parks and
gardens and for fire-fighting purposes. Surface detention ponds have been modified to store
stormwater for later re-use in Florida (Bradner and Wanielista, 1992; Wanielista and Bradner,
1992). This is in addition to employing the ponds for flood control and for improving the
amenity value of the water. Other uses could be for groundwater recharge, water storage during
the rainy season and subsequent withdrawal in the dry season. Groundwater recharge can also be
used to prevent seawater intrusion in coastal areas subject to heavy groundwater withdrawal in
excess of natural replenishment by precipitation.
2.2 STRUCTURAL OR TREATMENT BEST MANAGEMENT PRACTICES
Structural BMPs are used to treat the stormwater either at the point of generation or the point of
discharge to either the storm sewer system or to receiving waters. The selection and successful
design of selected structural BMPs for stormwater quality enhancement is the cornerstone of
stormwater management in newly developing and redeveloping urban areas. Structural BMPs
require commitment of resources for initial construction and continuing operation and
maintenance. Structural BMPs can be grouped into several general categories; however, the
distinction between BMP types and the terminology used to group structural BMPs is an area that
requires standardization. For purposes of this paper, and to be consistent with the definitions and
terminology used in the ASCE National Stormwater Database and the OW menu of BMPs,
structural BMPs for urban stormwater management have been grouped and defined as shown in
Table 2-2. (Schueler, 1987; WEF and ASCE, 1998; U.S. EPA, 1999b; NYSDEC, 2001).
Table 2-2. Structural or Treatment Best Management Practices for Urban Stormwater
Major Categories
Structural BMPs
2.2.1 Ponds
<
<
<
Dry Detention Ponds
Dry-Extended Detention Ponds
Wet (Retention) Ponds
2.2.2 Stormwater Wetlands
<
Constructed Wetlands
2 - 21
2.2.3 Vegetative Biofilters
<
<
<
Grass Swales (Wet/Dry)
Filter Strip/Buffer
Bioretention Cells
2.2.4 Infiltration Practices
<
<
<
Infiltration Trench
Infiltration Basin
Porous Pavement
2.2.5 Sand and Organic
Filters
<
<
<
<
Surface Sand Filter
Perimeter Filter
Media Filter
Underground Filter
2.2.6 Technology Options
and Others
<
<
<
Water Quality Inlets
Multi-Chambered Treatment Train
Vortex Separation/Continuous Deflection Systems
(WEF and ASCE, 1998; U.S. EPA, 2001a)
2.2.1 Stormwater Ponds
Stormwater ponds refer to practices that have either a permanent pool of water or a combination
of permanent pool and extended detention capable of treating the water quality volume (WQv),
which is the storage needed to capture and treat the runoff from 90% of average annual rainfall.
Treatment of the WQv shall be provided at all developments where stormwater management is
required.; a more detailed of the same follows in Chapter 3. Stormwater ponds can either be a
detention system (dry/extended detention ponds) or a retention system (wet ponds). Detention
systems capture a volume of runoff and temporarily retain that volume for subsequent release and
do not retain a significant permanent pool of water between runoff events. Retention systems, on
the other hand, capture a volume of runoff and retain that volume until it is displaced in part or in
total by the next runoff event, thereby maintaining a significant permanent pool of water between
runoff events (NYSDEC, 2001).
2.2.1.1 Dry Detention Ponds
Dry ponds, also called detention ponds, are stormwater basins designed to intercept a volume of
stormwater runoff and temporarily impound the water for gradual release to the receiving stream
or storm sewer system. They are usually designed to reduce the peak flow resulting from a
selected design storm (e.g., a 10-yr storm) to the pre-development level to prevent downstream
flooding. These typically on-line, end-of-pipe BMPs are designed to completely empty out
between runoff events, and therefore mainly provide runoff control as opposed to water quality
control (U.S. EPA, 1999b; Metro Council, 2001). Dry ponds are not very effective in removing
pollutants, especially particulates, due to the short detention times; particulates that settle to the
bottom of the pond are easily re-suspended by subsequent runoff. Dry ponds should therefore be
treated as practices used to reduce the peak discharges of stormwater to receiving streams, limit
2 - 22
downstream flooding and provide some degree of channel protection.
Most dry ponds are designed to empty in a time period of less than 24 h and can limit
downstream scour and loss of aquatic habitat by reducing the peak flow rate and energy of
stormwater discharges to the receiving stream. Typically, dry ponds are designed so that release
rates are comparable to pre-development flow rates, and their use is largely confined to retrofit
situations and as a part of an overall treatment-train approach (Clar et al., 2003).
2.2.1.2 Dry, Extended-detention (ED) Ponds
The outlet structure of a dry pond can be modified in such a way to provide a “retention outlet”
that is sized for slow release of the runoff from a designated “BMP storm”; a BMP storm is a
small and frequent storm, such as the 1-yr storm, which is prescribed by regulations or
ordinances as the BMP design storm (Yu and Nawang, 1993). ED ponds temporarily detain a
portion of urban runoff for up to 48 h (a 24 h limit is more common) after a storm, using a fixed
orifice to regulate outflow at a specified rate, allowing solids and associated pollutants the
required time to settle out. The ED ponds are normally "dry" between storm events, do not have
any permanent standing water and typically are composed of two stages: an upper stage, which
remains dry except for larger storms, and a lower stage, which is designed for typical storms.
The performance can be enhanced by using plunge pools near the inlet, a micropool at the outlet,
and an adjustable reverse-sloped pipe as the ED control device (orifice) (NVPDC, 1979; U.S.
EPA, 1993a). Temporary and most permanent ED ponds use a riser with an antivortex trash rack
on top to control trash.
Advantages:
Dry Ponds:
<
Can perform well in cold climates.
<
Limit downstream scour and loss of aquatic habitat by reducing the peak flow rate and
energy of stormwater discharges to receiving streams.
<
Properly designed ponds could be used as recreational areas when not in frequent use.
ED Ponds:
<
ED ponds are some of the best facilities for treating spring and winter runoff, as
compared to dry ponds; ponds without ED may have minimal storage above the ice
surface; therefore, treatment could be bypassed.
<
ED ponds are very effective in controlling peak discharges, which is an important factor
in reducing downstream streambank erosion and sediment loads.
<
These BMPs are good retrofitting options for existing dry basins and control both
stormwater quality and quantity.
<
ED ponds that include a dead storage pool can remove significant levels of sediment and
sorbed pollutants, and could provide excellent streambank erosion protection and
stormwater treatment when used in combination with other structural practices or when
retrofitted with permanent pools.
2 - 23
Limitations:
Dry Ponds:
<
Dry ponds are not a suitable option for drainage areas of less than 10 acres.
<
Dry ponds have a high potential for clogging outlets and sediment resuspension between
storm events if improperly maintained, and provide only marginal removal of stormwater
pollutants.
ED Ponds:
<
Erosion and resuspension of sediments may occur in the pond if the upper stage is not
properly vegetated.
<
Discharges from ponds may consist of warm water, hence, their use must be limited in
areas where warm water discharges will adversely affect a cold-water fishery.
2.2.1.3 Wet (Retention) Ponds
Wet ponds, also known as retention ponds, are designed to intercept a volume of stormwater
runoff and provide storage and treatment of this runoff volume. Water in the pond above the
permanent pool level, is the volume available for storage (WQv), that is displaced in part or
completely by the runoff volume from subsequent runoff events. Wet ponds have a capacity
greater than the permanent pond volume, which permits storage of the influent stormwater runoff
and controlled release of the mixed influent and permanent pond water (Field and Sullivan,
2003). Properly designed and maintained wet ponds can be extremely effective BMPs, providing
both water quality improvements and quantity control, in addition to providing aesthetic value
and aquatic and terrestrial habitat for a variety of plants and animals (U.S. EPA, 1999a). Wet
ponds allow particulate pollutants to settle out and dissolved pollutants to be removed by
biological uptake or other decay processes (Yu and Nawang, 1993).
Advantages:
<
Properly designed, constructed and maintained wet ponds can provide substantial
aesthetic/recreational value and wildlife and wetlands habitat.
<
The presence of a permanent wet pool helps provide significant water quality
improvement across a relatively broad spectrum of constituents, including dissolved
nutrients.
<
Widespread application with sufficient capture volume can provide significant control of
channel erosion and enlargement caused by changes to flow frequency relationships that
result from the increase of impervious cover in a watershed.
Limitations:
<
Wet ponds require relatively large land area and are not suited for drainage areas smaller
than 10 acres.
<
They cannot be placed on steep or unstable slopes and require a base flow or
supplemental flow in order to maintain the water level.
<
Improper design and irregular maintenance may result in stratification and anoxic
2 - 24
conditions that can promote the release of metals and nutrients from the trapped
sediments (CASQA, 2003).
2.2.2. Stormwater Wetlands
Wetlands have long been used in the final treatment of municipal wastewater, and in the last
decade have been increasingly used as a stormwater BMP option (Field and Sullivan, 2003).
Wetlands remove pollutants through sedimentation, plant uptake, microbial decomposition,
sorption, filtration, and exchange capacity, and can be natural, modified natural, or constructed
(Metro Council, 2001).
2.2.2.1 Constructed Wetlands
Constructed wetlands are engineered systems designed to simulate the water quality
improvement functions of natural wetlands to treat and contain surface water runoff pollutants
and decrease loadings to surface waters. Constructed wetland systems are similar to detention
and retention systems, except that a major portion of the BMP water surface area (in pond
systems) or bottom (in meadow-type systems) contains wetland vegetation and this system also
includes wetland channels (NYSDEC, 2001).
Constructed wetland systems incorporate the natural functions of wetlands to aid in pollutant
removal from stormwater and also control stormwater quantity by providing a significant volume
of ponded water above the permanent pool elevation. However, constructed wetlands treating
urban runoff differ from artificial wetlands created to comply with mitigation requirements in
that they do not replicate all of the ecological functions of natural wetlands. Enhanced designs
may include a forebay, complex microtopography, and pondscaping with multiple species of
wetland trees, shrubs, and plants (CASQA, 2003). Incorporating a sediment forebay or some
other pretreatment provision into the wetland design allows for the removal of coarse sediments,
which may otherwise degrade the performance of the system.
The use of stormwater wetlands is limited by a number of site constraints, including soil types,
depth to groundwater, contributing drainage area and available land area. Constructed wetlands
are especially appropriate where groundwater levels are close to the surface as groundwater can
supply the necessary water required for sustaining the wetland system. It has been observed that
medium-fine texture soils such as loams and silt loams are best to establish vegetation, retain
surface water, permit groundwater discharge and capture pollutants (Metro Council, 2001).
Advantages:
<
Flood attenuation, reduction of peak discharges and improvements in downstream water
quality.
<
Settlement of particulate pollutants, reduction of oxygen-demanding substances and
bacteria from urban runoff, and biological uptake of pollutants by wetland plant species.
<
Enhancement of vegetation diversity and wildlife habitat in urban areas as well as
2 - 25
aesthetic enhancement and valuable addition to community green space.
Limitations:
<
Maintenance of wetland vegetation may be difficult under a variety of flow conditions.
<
Pollutant removal efficiencies may be lower than anticipated until the establishment of
wetland vegetation.
<
The release of nutrients in the fall could have an impact on receiving water quality.
<
Wetlands may act as a heat sink and discharge warmer water to downstream water bodies.
2.2.3 Vegetated Systems (Biofilters)
Vegetated systems are practices explicitly designed to capture and treat the full WQv within dry
or wet cells formed by check dams or other means and include systems designed to convey and
treat either shallow flow (swales) or sheetflow (filter strips) runoff. These BMPs are commonly
referred to as "biofilters," since the grasses and vegetation "filter" the stormwater as it flows over
them (U.S. EPA, 1999b). A certain degree of treatment, storage and infiltration could be
provided by conveying stormwater runoff in vegetated systems, which helps to reduce the overall
volume of stormwater runoff generated from a particular drainage area; this makes open channel
vegetated systems a better alternative to traditional curb-and-gutter and storm sewer conveyance
systems (CASQA, 2003). While biofiltration swales are vegetated channels that receive directed
flow and convey storm water, biofiltration strips, also known as vegetated buffer strips, are
vegetated sections of land over which stormwater flows as overland sheet flow (Caltrans, 2002).
Pollutants are removed by filtration through the grass, sedimentation, adsorption to soil particles,
and infiltration through the soil (Field and Sullivan, 2003). Strips and swales are mainly
effective at removing debris and solid particles, although some dissolved constituents are
removed by adsorption onto the soil (Metro Council, 2001).
2.2.3.1 Grass Swales (Wet, Dry)
A grassed swale is an infiltration/filtration method that is usually used to provide pretreatment
before stormwater runoff is discharged to treatment systems. Grassed swales are typically
shallow, vegetated, man-made conveyance channels designed so that the bottom elevation is
above the water table to allow runoff to infiltrate into ground water. The vegetation or turf
covering the side slopes and channel bottom collect and slowly convey runoff to downstream
discharge points (CASQA, 2003). Swales are designed to treat stormwater runoff through
filtering by the vegetation in the channel, filtering through a subsoil matrix, and/or infiltration
into the underlying soils, during which they trap pollutants, promote infiltration, and reduce flow
velocities. Swales can be either wet or dry; while dry swales are used in areas where standing
water is not desired, such as residential areas, wet swales can be used where standing water does
not create a nuisance problem and where the groundwater is close enough to the surface to
maintain the permanent pool in inter-event periods. The added benefit of wet swales is the
ability to include in them a range of wetland vegetation to aid in pollutant removal (U.S. EPA,
1999b).
2 - 26
2.2.3.1.1 Dry Swales
The dry swale is a type of open vegetated channel used to treat and attenuate the WQv of
stormwater runoff, as well as convey excess stormwater downstream. The entire WQv of a given
storm is temporarily held in a pool or series of pools created by permanent checkdams or
ditchblocks. This holding time serves to settle pollutants, especially sediments. Typically
located in a drainage easement at the back side of a residential lot or along roadsides in place of
curbs and gutters, dry swales are good options in residential settings, as they discourage
long-standing water. This makes it possible to mow the area shortly after a rainfall event.
Stormwater treated by the soil bed flows into an underdrain system that conveys treated
stormwater back to the storm drain system (Metro Council, 2001).
Advantages:
<
Good option for small-area stormwater retrofits, replacing existing drainage ditches; for
residential or institutional areas of low to moderate density, and may also be used in
parking lots to break up impervious areas.
<
Linear nature of the design works well for treating highway or residential road runoff.
<
Rapid de-watering and shallow slopes that are easy to mow.
<
Since runoff ponds for only a short time in a dry swale, water temperatures do not
significantly increase, making this an appropriate practice for use in watersheds with
cold-water trout streams.
Limitations:
<
Individual dry swales can treat only a small area and may not be applicable to sites with
many driveway culverts or extensive sidewalk systems; roadside dry swales are subject to
damage from off-street parking and snow removal (Metro Council, 2001).
<
Do not appear to be effective in reducing bacteria levels in stormwater, and appear to
remove only modest amounts of phosphorus, with relatively few studies known to gauge
their effectiveness.
2.2.3.1.2 Wet Swale
The wet swale, also called a grassed open channel, consists of a broad open channel capable of
temporarily storing water, and unlike the dry swale, does not have an underlying filtering bed.
The wet swale is constructed directly within existing soils and may or may not intersect the water
table. Water quality treatment mechanisms rely primarily on settling of suspended solids,
adsorption, and microbial breakdown of pollutants. Like the dry swale, the entire water quality
treatment volume is stored and retained within a series of cells in the channel, formed by berms
or checkdams. Wet swales also reduce the velocity of stormwater runoff and may promote
infiltration (Metro Council, 2001). A wet swale is a suitable option when the water table is
located very close to the surface, with the result that swale soils often become fully saturated, or
have standing water all or part of the year after the channel has been excavated. Thus a wet
swale essentially acts as a very long and linear shallow wetland treatment system. Pollutant
removal rates in some cases may be enhanced by planting emergent wetland plant species in
2 - 27
these cells (CASQA, 2003).
Advantages:
<
Control peak discharges by reducing runoff velocity and promoting infiltration and
provide effective pretreatment for BMPs in series by trapping, filtering and infiltrating
pollutants.
<
Convey water in properly protected channels and divert water around potential pollutant
sources.
<
Provide water quality treatment by sedimentation and biological uptake and enhance
biological diversity and create beneficial habitat between upland and surface waters.
Limitations:
<
Impractical in areas with very flat grades, steep topography, wet or poorly drained soils.
<
Possibility of erosion when flow volumes and/or velocities are high during storm events.
<
Area requirements can be excessive for highly developed sites.
<
Roadside wet swales become less feasible with an increase in the number of driveway
entrances requiring culverts.
2.2.3.2 Vegetative Filter Strips (VFS)
According to the U.S. EPA (1993a), a vegetative buffer area or strip is defined in many cases as a
"permanent, maintained strip of planted or indigenous vegetation located between nonpoint
sources of pollution and receiving water bodies for the purpose of removing or mitigating the
effects of diffuse source pollutants such as nutrients, pesticides, sediments, and suspended
solids." These are vegetated strips of land that act as “buffers” by accepting storm runoff as
overland sheet flow from upstream developments and providing similar treatment potential
mechanisms to that of swales, prior to discharge of the storm runoff to the storm drainage system
(Field and Sullivan, 2003). They may closely resemble many natural ecotones, such as grassy
meadows or riparian forests, and their dense vegetative cover facilitates sediment attenuation and
pollutant removal. Vegetative filter strips (VFS) are frequently planted with turf grass; however,
alternatives that adopt any natural vegetated form such as meadows or small forest may be used.
Originally used as an agricultural practice, filter strips are now evolving as an urban practice
(CASQA, 2003). However, VFS, unlike grassed swales, are effective only for overland sheet
flow and provide little treatment for concentrated flows. Therefore, they are more suitable for
use in agriculture and low-density development, as well as other situations where runoff tends
not to be concentrated. VFS are often used as pretreatment for other structural practices, such as
infiltration basins and infiltration trenches, and land grading and/or a level spreader can be used
to create a uniformly sloping area that distributes the runoff evenly across the filter strip (U.S.
EPA, 1993a). They can also be used in combination with riparian buffers in treating sheet flows
and in stabilizing drainage channel banks and stream banks. In semi-arid climates, grass filter
strips may need to be irrigated to maintain a dense stand of vegetation and to prevent export of
unstabilized soil (Metro Council, 2001).
2 - 28
Advantages:
<
VFS work well in residential areas, where they provide open space for recreation
activities, help maintain riparian zones along streams, reduce streambank erosion and
provide animal habitat.
<
VFS can be useful as sediment filters during construction, which in some cases, may
require only the preservation of an appropriately located area of existing vegetation.
<
Flow characteristics and vegetation type and density can be closely controlled to
maximize BMP effectiveness.
Limitations:
<
VFS cannot treat a very large drainage area and they require a thick vegetative cover to
ensure proper functioning.
<
This BMP does not provide significant attenuation of the increased volume and flow rate
of runoff during intense rain events and is best implemented as one of a series of
stormwater BMPs.
<
Vegetative buffers may not provide treatment for dissolved constituents except to the
extent that flows across the vegetated surface are infiltrated into the soil profile.
<
VFS is not recommended to treat highly contaminated "hotspot" runoff, since infiltration
could result in groundwater pollution and damage to vegetation.
2.2.3.3 Bioretention
Bioretention systems are designed to mimic the functions of a natural forest ecosystem for
treating stormwater runoff. According to one definition, “bioretention systems are a variation of
a surface sand filter, where the sand filtration media is replaced with a planted soil bed” (U.S.
EPA, 1999b). Another approach considers bioretention as a “concept that uses biologic activity
(plants and microbes) to filter/clean stormwater by being incorporated into different kinds of
infiltration and filtration BMP designs such as infiltration basins, rainwater gardens and surface
sand filters” (Metro Council, 2001). This white paper views bioretention as a BMP that
functions as a soil and plant-based filtration device for stormwater management and removes
pollutants through a variety of physical, biological and chemical treatment processes. These
facilities normally consist of a grass buffer strip, sand bed, ponding area, organic layer or mulch
layer, planting soil and plants.
In general, bioretention systems can be described as shallow, landscaped depressions commonly
located in parking lot islands or within small pockets in residential areas that receive stormwater
runoff. Stormwater flows into the bioretention area, ponds on the surface, and gradually
infiltrates into the soil bed. A number of processes including adsorption, filtration, volatilization,
ion exchange and decomposition are responsible for pollutant removal (PGDER, 1993). Filtered
runoff can either be allowed to infiltrate into the surrounding soil (functioning as an infiltration
basin or rainwater garden), or collected by an underdrain system and discharged into the storm
sewer system or directly to receiving waters (functioning like a surface sand filter). Runoff from
larger storms is generally diverted past the area to the storm drain system. The bioretention
2 - 29
system has been used as a stormwater BMP since 1992. In addition to Prince George's County,
MD and Alexandria, VA, bioretention has been used successfully in urban and suburban areas in
Montgomery County, MD; Baltimore County, MD; Chesterfield County, VA; Prince William
County, VA; Smith Mountain Lake State Park, VA; and Cary, NC (CASQA, 2003).
A bioretention system includes the following components (Metro Council, 2001):
1. Grass Buffer Strips: Runoff enters the bioretention area as sheet flow through the grass buffer
strips where the buffers reduce runoff velocity and filter particulates from the runoff.
2. Ponding Area: The ponding area provides for surface storage of storm runoff before it filters
through the soil bed. It also allows for the evaporation of ponded water as well as the settling of
sediments in the runoff.
3. Organic Mulch Layer: The organic mulch layer protects the soil bed from erosion, retains
moisture in the plant root zone, provides a medium for biological growth and decomposition of
organic matter and provides some filtration of pollutants.
4. Planting Soil Bed: The planting soil bed provides water and nutrients to support plant life in
the bioretention system. Stormwater filters through the planting soil bed where pollutants are
removed by filtration, plant uptake, adsorption and biological degradation.
5. Sand Bed: The sand bed underlies the planting soil bed and allows water to drain from the
planting soil bed through the sand bed and into the surrounding soil. It also provides additional
filtration and allows for aeration of the planting soil bed.
6. Plants: Plants are important components of the bioretention system, and remove water through
evapotranspiration and pollutants and nutrients through uptake. The plant species selected are
designed to replicate a forested ecosystem and to survive stresses such as frequent periods of
inundation during runoff events and drying during inter-event periods.
Advantages:
<
Properly designed and maintained bioretention systems are more likely to be aesthetically
pleasing than other types of filtration or infiltration systems due to the presence of plants.
<
Layout of bioretention facilities can be very flexible, and the selection of plant species
can provide for a wide variety of landscape designs.
<
Ideally suited to highly impervious areas, such as parking lots, and can be applied in
many different climates and geologic environments, with some minor design
modifications.
<
Reduce the volume of runoff from a drainage area and can be very effective for removing
fine sediments, trace metals, nutrients, bacteria and organics (U.S. EPA, 1999a).
Limitations:
<
Bioretention as a BMP is not recommended for areas with slopes greater than 20% or
2 - 30
<
<
<
where mature tree removal would be required since clogging may result, particularly if
the facility receives runoff with high sediment loads.
Bioretention is not a suitable BMP in locations where the water table is within 6 feet of
the ground surface and where the surrounding soil stratum is unstable.
Susceptible to clogging by sediment; therefore pretreatment is a necessary part of design.
In cold climates, the soil may freeze, preventing runoff from infiltrating into the planting
soil.
By design, bioretention BMPs have the potential to create very attractive habitats for
mosquitoes and other vectors because of their highly organic, often heavily vegetated
areas mixed with shallow water.
2.2.4 Infiltration Systems
Infiltration practices have a high potential of controlling stormwater runoff by disposal at a local
site level. An infiltration BMP is designed to capture a volume of runoff, retain it and allow it to
infiltrate into the ground. There are a number of advantages and disadvantages associated with
infiltration practices (Schueler, 1987). Advantages include the control of both the quality and
quantity of stormwater. Water quantity is controlled by controlling surface runoff infiltration
into the underlying soil. This reduces the volume of water discharged to receiving streams and
thereby some of the potential impacts caused by excess flows, as well as reducing increased
pollutant concentrations in the receiving stream. They can be designed to capture a volume of
stormwater and infiltrate this volume into the ground over a period of several hours or even days,
thereby maximizing the infiltrating capacity of the BMP. Secondary benefits include increased
recharge of underlying aquifers, increasing base flow levels of nearby streams, and water quality
treatment. Pollutant removal occurs as water percolates through the various soil layers and
particulates are filtered out in addition to soil microbial degradation of organic pollutants
contained in the infiltrated stormwater. The disadvantages are just as numerous. Infiltration may
not be appropriate in areas where groundwater is a primary source of drinking water due to this
method’s potential for contaminant migration. This holds true when runoff is from a commercial
or residential area with a higher potential for metal or organic contamination. Also, the
performance is limited in areas with poorly permeable soils, and these BMPs can experience
reduced infiltrating capacity and clogging due to excessive sediment accumulation.
2.2.4.1 Infiltration Basins
Infiltration basins are typically off-line, end-of-pipe BMPs, and are designed to intercept only a
certain volume of runoff. A flow splitter or weir is usually used to divert runoff from a storm
sewer system into the infiltration basin. The basin may or may not be lined with plants.
However, vegetated infiltration systems help to prevent migration of pollutants, and the roots of
the vegetation can increase the permeability of the soils, thereby increasing the efficiency of the
basin. The main purpose of this BMP is to transform a surface water flow into a groundwater
flow and to remove pollutants through mechanisms such as filtration, adsorption and biological
conversion as the water percolates through the underlying soils. Infiltration basins should drain
2 - 31
within 72 h to maintain aerobic conditions that favor bacteria that aid in pollutant removal, and to
ensure that the basin is ready to receive the next storm (Schueler, 1987; U.S. EPA, 1993b).
Infiltration basins are inappropriate for areas that contribute high concentrations of sediment or
suspended solids without adequate treatment. Pretreatment, such as grit chambers, swales with
check dams, filter strips, or a sedimentation basin should be a fundamental component of any
BMP system relying on infiltration. Runoff entering the basin is pretreated to remove coarse
sediment that may clog the surface soil pores on the basin floor. Concentrated runoff should
flow through a sediment trap, or a filter strip may be used for sheet flow.
Infiltration basins are dry ponds constructed to allow infiltration to occur simultaneously with
other treatment processes. The operating characteristics of infiltration basins are essentially the
same as for dry detention ponds (Field and Sullivan, 2003), except a few significant exceptions
(Clar et al., 2003):
<
Infiltration basins also remove dissolved and colloidal solids in the volume of infiltrated
water, whereas dry ED ponds can remove only the fraction of colloidal solids sorbed to
settleable solids.
<
The settling velocities of sediment particles and particulate (settleable) chemicals are
increased by a value equal to the infiltration rate in the basin. The impact would be more
for clay (colloidal) sized particles than for silt, sand, and small or large aggregates.
<
Infiltration practices differ from typical dry basins because they contribute to groundwater
recharge, thereby providing an additional element of performance.
<
By providing volume control, these BMPs can effectively address the increased frequency
and duration of peak flows and can provide downstream channel protection.
<
The mode of operation being infiltration of runoff into subsurface soils, infiltration basins
are able to preclude the thermal impacts issues associated with detention, ED and wet
ponds.
Advantages:
<
The principal benefit of infiltration basins is the approximation of pre-development
hydrology during which a significant portion of the average annual rainfall runoff
infiltrates and evaporates rather than flushing directly to creeks.
<
Infiltration basins can be useful for controlling channel forming (erosion) and high
frequency (generally less than the 2-yr) flood events by adequately sizing WQv.
<
Provides groundwater recharge and baseflow in nearby streams, reduces local and
downstream flooding and protects streambank integrity.
<
Can be very effective for removing fine sediment, trace metals, nutrients, bacteria and
oxygen-demanding substances.
Limitations:
<
Potentially high failure rates due to improper siting, design and lack of maintenance,
especially if pretreatment is not incorporated into the design.
<
Not appropriate for treating significant loads of sediment and other pollutants due to the
2 - 32
<
potential for clogging.
Not suitable on fill sites or steep slopes and a risk of groundwater contamination in very
coarse soils.
2.2.4.2 Infiltration Trenches
Infiltration trenches are shallow (3 to 12 ft) excavated ditches that are lined with filter fabric and
filled with stone to create underground reservoirs for stormwater runoff from a specific design
storm. Urban runoff diverted into the trench gradually infiltrates through the bottom and sides of
the trench into the subsoil and eventually into the groundwater over a period of days (Metro
Council, 2001). Infiltration trenches are typically implemented at the ground surface to intercept
overland flows. Runoff can be captured by depressing the trench surface or by placing a berm at
the down gradient side of the trench (Schueler, 1987). Design variations include dry wells, pits
designed to control small volumes of runoff (such as the runoff from a rooftop), and enhanced
infiltration trenches, which are equipped with extensive pretreatment systems to remove
sediment and oil. Depending on the quality of the runoff, pretreatment will generally be
necessary to lower the failure rate of the trench. More expensive than pond systems in terms of
cost per unit of runoff treated, infiltration trenches are best suited for drainage areas of less than 5
to 10 acres, or where ponds cannot be applied (Schueler, 1992).
The design storm for an infiltration trench is typically a frequent, small storm such as the 1-yr
event, that provides treatment for the "first flush" of stormwater runoff. Hence they are
frequently used in combination with other BMPs such as a detention basin to control peak
hydraulic flows. Infiltration trenches provide total peak discharge, runoff volume and water
quality control for all storm events equal to or less than the design storm. This infiltration
reduces the runoff volume, removes many pollutants and provides stream baseflow and
groundwater recharge (CASQA, 2003).
Advantages:
<
Provides 100% reduction in the load discharged to surface waters.
<
The principal benefit of infiltration basins is the approximation of pre-development
hydrology during which a significant portion of the average annual rainfall runoff is
infiltrated and evaporated rather than flushed directly to creeks.
<
If the WQv is adequately sized, infiltration basins can be useful for providing control of
channel forming (erosion) and high frequency (generally less than the 2-yr) flood events.
<
As an underground BMP, trenches are unobtrusive and have little impact on site
aesthetics.
Limitations:
<
High failure rate if soil and subsurface conditions are unsuitable.
<
Unsuitable on fill sites or steep slopes and a risk of groundwater contamination in very
coarse soils.
<
Upstream drainage area must be completely stabilized before construction.
2 - 33
2.2.4.3 Porous Pavement
Porous pavement is an infiltration system in which stormwater runoff infiltrates into the ground
through a permeable layer of pavement or other stabilized permeable surface. The permeable
surface can include porous asphalt, porous concrete, modular perforated concrete block, cobble
pavers with porous joints, gaps or reinforced/stabilized turf (Florida DER, 1988). Concrete grid
pavement consists of concrete blocks with regularly inter-dispersed void areas with pervious
materials, such as gravel, sand, or grass. The blocks are typically placed on a sand or gravel base
and designed to provide a load-bearing surface that is adequate to support vehicles, while
allowing infiltration of surface water into the underlying soil. Porous pavement provides an
alternative to conventional pavement and reduces much of the need for urban runoff drainage
conveyance and treatment offsite. Instead, runoff is diverted through a porous asphalt layer into
an underground stone reservoir before gradually exfiltrating out of the stone reservoir into the
subsoil (Field and Sullivan, 2003). Porous pavements tend to clog with fine sediments and need
to be subjected to periodic vacuum-type street sweeping (WA DOE, 1991). Modular pavers,
grassed parking areas, and permeable pavements may also be employed to reduce runoff volumes
and trap vehicle-generated pollutants (U.S. EPA, 1993a); however, care should be taken when
selecting such alternatives. The potential for groundwater contamination, compaction, or
clogging due to sedimentation should be evaluated during the selection process. These practices
should be selected only in cases where proper operation and maintenance can be guaranteed due
to high failure rates without proper upkeep.
There has been mixed observations on the performance of porous pavement systems. The
performance of porous asphalt has been historically very poor in the mid-Atlantic region.
However, many of these failures can be attributed to lack of proper erosion and sediment controls
during construction or lack of contractor experience with installation of porous pavement
systems. Yet, studies on porous concrete systems in use in Florida show that they have
performed very well (U.S. EPA, 1999b).
Advantages:
<
When properly designed and maintained, porous pavement systems can be an effective
means of managing stormwater runoff, and particularly useful for overflow parking areas
that are not used on a daily basis.
<
Especially useful for driveways and streets and in residential areas, and in commercial
parking lots.
Limitations:
<
Not effective in areas that receive runoff with high amounts of sediments due to the
tendency of the pores to clog.
<
Require maintenance, including periodic vacuuming or jet-washing to remove sediment
from the pores.
2 - 34
2.2.5 Filtration Systems
A filtration system is a device that uses a combination of a granular filtration media such as
sand, gravel, organic material (e.g., peat, compost), membrane, or other acceptable treatment
media to remove pollutants in stormwater runoff. Sand filters are a self-contained bed of sand to
which the first flush of runoff is diverted. The runoff percolates through the sand, where
colloidal and particulate materials are strained out by the cake of solids that forms, or is placed
on the surface of the media. Water leaving the filter is collected in underground pipes and
returned to the stream or channel. A layer of peat, limestone, and/or topsoil may be added to
improve removal efficiency (CASQA, 2003). Detention time is typically 4 to 6 hs in a filter, and
sediment-trapping structures are typically used to prevent premature clogging of the filter media
(NVPDC and ESI, 1992). Filtration systems are primarily water quality control BMPs designed
to remove particulate pollutants in stormwater; quantity control can also be included by
providing additional storage volume in a pond or basin, or vertical storage volume above the
filter bed, or by allowing water to temporarily pond in parking lots or other areas before being
discharged to the filter. Sand filters have been demonstrated to be effective in removing many of
the common pollutants found in urban stormwater runoff, especially the particulate pollutants.
While they seem to have a moderate level of bacterial removal, they have not been effective at
removing total dissolved solids and nitrate-nitrogen (Metro Council, 2001). Media filters can be
successfully employed for stormwater management in (i) small sites such as parking lots and
developments; (ii) areas with high pollution potential such as industrial areas; or, (iii) highly
urbanized areas, where land availability or costs preclude the use of other BMP types (U.S. EPA,
1999b). Filters should be placed off-line, (i.e., the water quality volume is diverted to the BMP,
while any flows in excess of this volume are bypassed) and are sometimes designed to intercept
and treat only the first flush and bypass larger storm flows (CASQA, 2003).
Commonly used types of filters include: (i) surface sand filters such as the "Austin" sand filter;
and (ii) underground vault filters such as the "DC" sand filter and the "Delaware" sand filter.
Design variations of these basic types exist in addition to several proprietary filtering systems;
some use specialized filter media made from materials such as leaf compost (Metro Council,
2001). This is in addition to a number of variations in the types of filtration media used
described earlier. These designs may also incorporate additional features such as a layer of filter
cloth or a plastic screen, gravel layer, peat layer, compost layer, layer of peat or a peat/sand
mixture.
2.2.5.1 Surface Sand Filter
The surface sand filter was initially developed in Florida in 1981 for sites that could not infiltrate
runoff or were too small for effective use of detention systems, with further development of filter
technology by the city of Austin, Texas. A surface sand filter consists of a pretreatment basin,
water storage reservoir, flow spreader, sand and underdrain piping. A basin liner may also be
needed if the treated runoff is not to be allowed to infiltrate into the soil underlying the filtration
basin because of groundwater concerns. This system usually incorporates two basins: The first
2 - 35
is a sedimentation basin, dry or wet, where runoff enters and is removed of coarse particulates by
gravity settling. Water then flows over a weir or through a riser into the filter basin, which
consists of sand with a gravel and perforated pipe underdrain system to capture the treated water.
The surface of the filter bed may be planted with grass. Additional storage volume is provided
above the filter bed to increase the volume of water that can be temporarily ponded in the system
prior to infiltration. This two basin configuration helps to limit premature clogging of the filter
bed due to excessive sediment loading. There are several variations of this design, Austin uses
two variations, partial, and full sedimentation filtration systems (City of Austin, 1988).
2.2.5.2 Underground Vault Sand Filter
An underground filter is similar to a surface filter except that the sand (or other media) and
underdrains are installed below grade in a vault. The underground vault sand filter was
developed by the District of Columbia in the 1980s. Its design incorporates three chambers. The
first chamber and the throat of the second contain a permanent pool of water and function as a
sedimentation chamber and an oil and grease and floatables trap, besides providing temporary
runoff storage. The two chambers are connected by a submerged opening or inverted elbow near
the bottom of the dividing wall, which provides a water seal that prevents the transfer of oil and
floatables to the second chamber containing the filter bed. During a storm event, water flows
through the opening into the second chamber and onto the filter bed. Additional runoff storage
volume is provided above the filter bed. Filtered water is collected by a gravel and perforated
pipe underdrain system and flows into the third chamber, which contains a clearwell and is
connected to the storm drain system. Overflow protection can be provided by placing the filter
off-line or by providing a weir at the top of the wall connecting the filter chamber with the
clearwell chamber to serve as an overflow. The schematic of an underground vault sand filter is
presented in Chapter 3 (Figure 3-11).
2.2.5.3 Perimeter Filter
A perimeter filter is an underground vault sand filter, and is referred to as a perimeter sand filter
because of its particular suitability for use around the perimeters of parking lots. This system,
developed in Delaware, and also known as the "Delaware Sand Filter", contains two chambers
and a clear well. Stormwater enters the first chamber, which serves as a sedimentation chamber.
Water then flows over a series of weirs and into the second chamber, which contains the filter
media. Additional storage volume is provided by water temporarily ponding in both chambers.
Filtered water is collected by a series of gravel and perforated pipe underdrains, and flows into a
clearwell that contains a connection to the storm drain system.
2.2.5.4 Others
In addition to the three basic filtering systems, in DC, Austin, Delaware, there are a number of
variations in use. A compound stormwater filtering system developed by the city of Alexandria,
VA incorporates an anoxic filtration zone in a permanently flooded gravel layer in the filter.
2 - 36
This anoxic zone aids in nitrogen removal by anoxic denitrification. Another configuration uses
an upflow anaerobic filter upstream of the sand filter to enhance phosphorus removal by
precipitating more iron on the sand filter. Yet another type employs organic materials such as
peat or compost combined with sand or other materials. Filters that use an organic filtration
media, such as peat or leaf compost, are useful in areas where additional nutrient or metal control
is desirable due to the adsorptive capacity, its non-exchange capability, and its ability to serve as
a medium for the growth of a variety of microorganisms. However, peat must be carefully
selected (fibric and/or hemic, not sapric, peat should be used) and the environmental concerns
regarding destroying peat bogs to obtain filtration media need to be addressed when other
technologies are available (Schueler, 1987; VA DCR, 1999).
Advantages:
<
Media filters can be used in high density urban sites with small drainage areas that are
completely impervious such as parking lots. They can be applicable to many areas that
are difficult to retrofit due to space limitations and used in sites where soil or
groundwater concerns do no support an infiltration BMP.
<
Applicable in small drainage areas of 1 to 10 acres; take up little space and can be used
on highly developed and steeply sloped sites.
<
Provide high removal efficiencies for total suspended solids (TSS).
Limitations:
<
Pretreatment required to prevent the filter media from clogging.
<
Not applicable in areas of high water tables and should not be used in areas where heavy
sediment loads are expected or in tributary areas that are not fully stabilized.
2.2.6 Technology O ptions and Others
This group of BMPs includes a variety of proprietary and miscellaneous systems that are used for
urban stormwater management and that do not seem to fall exclusively under any of the abovementioned categories. These include water quality inlets, hydrodynamic devices, filtration
devices, etc. Many of these systems are "drop-in" systems, and incorporate some combination of
filter media, hydrodynamic sediment removal, oil and grease removal, or screening to remove
pollutants from stormwater. There are also quite a number of proprietary devices in this
category, but this white paper does not focus on vendor-supplied systems and other proprietary
devices due to the lack of peer reviewed performance data for these systems.
2.2.6.1 Water Quality Inlets
Water Quality Inlets (WQI), also commonly called trapping catchbasins, oil/grit separators or
oil/water separators, consist of one or more chambers that promote sedimentation of coarse
materials and separation of free oil, as opposed to emulsified or dissolved oil, from stormwater.
Some WQIs also contain screens to help retain larger or floating debris, and many of the newer
designs also include a coalescing unit that helps promote oil/water separation. A typical WQI
2 - 37
consists of a sedimentation chamber, an oil separation chamber and a discharge chamber (U.S.
EPA, 1999a; Metro Council, 2001).
WQIs are underground retention systems designed to remove settleable solids, and there are
several design variations. The simplest form of design is a single-chambered urban runoff inlet
in which the bottom has been lowered to provide 2 to 4 ft of additional space between the outlet
pipe and the structure bottom for collection of sediment. Some WQIs include a second chamber
with a sand filter to provide additional removal of finer suspended solids by filtration; the first
chamber provides effective removal of coarse particles and helps prevent premature clogging of
the filter media. Other WQIs include an oil/grit separator. A typical oil/grit separator consists of
three chambers: The first chamber removes coarse material and debris; the second chamber
provides separation of oil, grease, and gasoline; and the third chamber provides safety relief
should blockage occur (NVPDC and ESI, 1992). WQIs typically capture the first portion of
runoff for treatment and are generally used for pretreatment before discharging to other BMPs
(Schueler et al., 1992).
Advantages:
<
WQIs can effectively trap trash, debris, oil and grease, and other floatables that would
otherwise be discharged to surface waters.
<
A properly designed and maintained WQI can serve as an effective BMP for reducing
hydrocarbon contamination and spills in receiving water sediments.
<
WQIs are generally recommended for drainage areas of 1 acre or less and are effective for
industrial hotspots that have the potential for petroleum-contaminated process washdown,
spills, and stormwater runoff.
Limitations:
<
WQIs have limited ability to separate dissolved or emulsified oil from runoff and are also
not very effective at removing nutrients and heavy metals, except where the metals
removal is directly associated with sediment removal.
<
High sediment loads can interfere with the ability of the WQI to effectively separate oil
and grease from the runoff and during high flow conditions, sediment residuals may be
resuspended and released from the WQI to surface waters.
<
WQIs generally provide limited hydraulic and residuals storage as a result of which they
do not provide substantial stormwater improvement.
<
Standing water in the devices can be a breeding ground for mosquitoes and other vectors,
and lack of maintenance often results in resuspension of settled pollutants.
2.2.6.2 Multi Chambered Treatment Train (MCTT)
A multiple treatment system uses two or more BMPs in series and the MCTT is an example of a
stormwater device that utilizes a combination of processes, especially pretreatment of stormwater
using sedimentation, followed by media filtration (Field and Sullivan, 2003). The MCTT was
developed during an U.S. EPA-sponsored research program on treatability of stormwater at
2 - 38
critical source areas, and areas designated as stormwater hotspots; typical locations include gas
stations, junkyards, bus barns, public works yards, car washes, oil change facilities, transmission
repair shops, auto repair facilities, fast-food restaurants, convenience stores etc., and other areas
where stormwater has a high probability of high concentrations of oils and other toxic organic
pollutants that are difficult to treat by other means (Pitt et al., 1999). The MCTT contains
aeration, sedimentation, sorption and sand-peat filtration, and is most suitable for use at relatively
small and isolated paved critical source areas (0.25 to 2.5 acre), where surface land is not
available for stormwater controls (Field and Sullivan, 2003). The MCTT consists of three
chambers:
<
a catchbasin or grit chamber for the removal of large particles and litter (screening),
<
a settling chamber for quiescent settling of fine settleable solids, and
<
a sand-peat moss "filter" for final polishing (removal of dissolved pollutants).
The collected runoff is first treated in a catchbasin chamber where larger particles are removed
by settling. The water then flows into a main settling chamber containing oil sorbent material
where it undergoes a much longer treatment period (24 to 72 h) to remove finer particles and
associated pollutants. The final chamber contains a mixed media filter material comprising equal
amounts of sand and peat. This chamber acts as a polishing "filter" to remove some of the
filterable toxicants from runoff by other processes, such as ion exchange and sorption. The
MCTT was designed to remove pollutants of a specific class of concern in stormwater:
particulates as small as a few µm and associated particulate bound toxicants, plus filterable
toxicants. Pilot and full scale test results showed substantial reductions in both particulate and
dissolved phases of stormwater toxicants as well as suspended solids. While substantial
reductions are observed in the concentration of pollutants and suspended solids in the main
settling chamber as well as the peat-sand chamber, the catchbasin/grit chamber did not provide
any significant improvements in water quality, although it is an important element in reducing
maintenance problems by trapping bulk material (GVSDD, 1999). Additional information on the
performance data of MCTTs is provided in (Pitt et al., 1999; Field and Sullivan, 2003).
Advantages:
The MCTT is capable of reducing a broad range of stormwater pollutants in particulate
fractions (as small as a few µm) and filterable fractions that cause substantial receiving
water problems.
<
Although unit construction cost appears to be expensive (construction costs estimated at
$15,000 to $30,000 per 0.1 ha drainage area plus 11 to 13% of construction costs for
operation and maintenance per yr (GVSDD, 1999), with right planning and
implementation, the MCTT has a high potential for cost-effective use as an integrated
component in watershed management programs designed to protect and enhance
receiving waters.
<
They can be used in areas where site conditions prevent the use of other BMPs such as
infiltration, wet ponds and wetlands and in space-limited areas because of the
underground design.
<
Relatively few concerns associated with aesthetics and safety.
<
2 - 39
Limitations:
<
The MCTT is designed for contaminant removal only and provides no significant
attenuation of peak flows or runoff volumes.
<
Lack of maintenance and undersized separators for the flows encountered.
<
Scouring of previously captured material is commonly observed.
2.2.6.3 Vortex Separators and Continuous Deflective Systems
Several solid/liquid separators use circular flow patterns to create inertial flow separation. One
general classification of these devices includes swirl and vortex separators that have been in use
for over 30 years to control combined sewer overflows (CSOs). Vortex separators, or swirl
concentrators, are gravity separators, and in principle are essentially wet vaults. The difference
between them and wet vaults, however, is that vortex separators are round rather than
rectangular, and water moves in them in a centrifugal fashion rather than a straight line before
exiting. By this approach, it is possible to obtain significant removal of suspended sediments and
attached pollutants with less space (U.S. EPA, 1977; Field et al., 1997). These systems were
originally developed to treat CSOs for the removal of coarse inorganic solids. They are
increasingly being used to control pollution from stormwater discharge (Konieek et al., 1996;
Field et al., 1997; CASQA, 2003) by reducing solids concentration in urban runoff (Lee et al.,
2003).
Hydrodynamic vortex separation is a proven technology with an established track record for
improving urban water quality (including CSOs, stormwater and wastewater) (Andoh and Saul,
2002). Hydrodynamic vortex separators are characterized by tangential flows into a cylindrical
vessel, which in turn creates a complex rotary regime and provides three main functions: flow
regulation, settleable solids concentration and the capture of floatables (U.S. EPA, 1977; Field et
al., 1997). Stormwater enters the vortex tangentially through an entry port approximately
halfway up the chamber wall so that it creates a swirling vortex flow pattern. The first flow field
allows the solids to settle out by gravity; the second vortex that is generated in the skirt causes
settleable solids to be concentrated at the bottom. The different configurations that have been
developed are differentiated by the nature and type of internal flow modifying components and
the location of inlets and outlets. The advantages of the vortex (or swirl) concentrator are that it
includes no moving parts or external power requirement and has a high hydraulic loading, which
results in compact size and low operational cost (Field and O'Connor, 1996). The successful
application of swirl and vortex technologies in a stormwater treatment system depends on the
following factors:
<
consistent and appropriate flow measurement, wastewater sampling and characterization
protocols,
<
appropriate data management technique, especially the calculation efficiency,
<
an understanding of swirl/vortex mechanisms, with realistic performance expectations,
and
<
appropriate application or placement in the treatment system (Field et al., 1997).
2 - 40
Continuous Deflective Separator (CDS) appears to have similar features to the vortex separator,
but the CDS design introduces a filtration mechanism for solids separation. The filtration
mechanism, when combined with circular flow action and particle sedimentation, increases
removal rates beyond vortex separators during times of high flows (Schwarz and Wells, 1999).
The CDS unit works by deflecting the inflow and associated pollutants away from the main flow
stream into a separation chamber. The chamber has a sump at the bottom and a screen in the
upper section. The screen acts to remove the gross pollutants allowing the filtered water to pass
through to a return system, while the floatable solids are kept in continuous motion on the water
surface by incoming flow in the chamber, thus keeping the solids in the chamber from blocking
the screen. With the heavy solids settling to the bottom of the sump, the CDS unit acts as a
continuous cleaning unit, since the solids do not become imbedded in the filter screen as in a
direct screening situation (Schwarz and Wells, 1999). The CDS mechanism is one of the new
methods for removing suspended solids from surface runoff, and studies show that screen sizing
has a significant influence on the sediment trapping efficiency of the CDS unit (Wong et al.,
1996).
Advantages:
<
May provide the desired performance in less space and less cost compared to other
BMPs.
<
May be more cost-effective pre-treatment devices than traditional wet or dry basins.
<
Mosquito control may be less of an issue than with traditional wet basins.
Limitations:
<
It is likely that vortex separators are not as effective as wet vaults at removing fine
sediments, on the order <50 to 100 µm in diameter.
<
Area served is limited by the capacity of the largest models.
<
The non-steady flows of stormwater decrease the efficiency of vortex separators from
what may be estimated or determined from testing under constant flow.
<
Do not remove dissolved pollutants; a loss of dissolved pollutants may occur as
accumulated organic matter decomposes in the units.
2.3 REMOVAL PROCESSES OCCURRING IN STRUCTURAL BMPS
The high population of pollutant sources in urbanized areas contributes large quantities of
pollutants that accumulate on streets, rooftops and other surfaces. During rainfall or snowmelt,
these pollutants are mobilized and transported into the storm drain system, where they are
conveyed and ultimately discharged to waterways. BMPs are implemented to remove these
pollutants from runoff in order to reduce the impacts to receiving waters. A wide range of
pollutants in urban runoff can be removed by properly designed, constructed and maintained
structural BMPs through a number of physical and biochemical processes. The U.S. EPA
identifies 115 organic compounds and 14 metals as "priority pollutants" in stormwater streams
that can be removed by six main processes (Scholze et al., 1993):
2 - 41
<
<
<
<
<
<
adsorption on suspended solids followed by particle sedimentation or filtration,
adsorption on an absorbing medium such as peat or other organic material,
volatilization (common during aeration processes),
photodegradation via photolysis,
biodegradation (aerobic and anaerobic)/biological nitrification/de-nitrification, and
phytoremediation/bioassimilation/biological uptake.
One or more of these treatment processes may occur in the treatment BMP systems to remove the
pollutants of concern in stormwater runoff (Table 2-3).
Table 2-3. Pollutant Removal Mechanisms in Common Urban Stormwater BMPs
Mechanism
Pollutants Affected
Promoted by
Sedimentation
Solids, BOD, pathogens,
COD (chemical oxygen
demand), P, N, metals
Low turbulence
Filtration
Solids, BOD, pathogens,
COD, P, N, metals
Fine, dense herbaceous plants
Sorption
Dissolved P, metals, synthetic High soil Al, Fe; high soil
organics
organics, circum-neutral pH
Oxidation
COD, petroleum
hydrocarbons, synthetic
organics
Aerobic conditions
Volatilization
volatile petroleum
hydrocarbons and synthetic
organics
High temperature and air
movement
Precipitation
Dissolved P, metals
High alkalinity
Biological Nitrification
NH3 - N
Dissolved oxygen > 2.0
mg/L, low toxics, temp. > 5-7
C, circum-neutral pH
Microbial Decomposition
BOD, COD, petroleum
hydrocarbons, synthetic
organics
High plant surface area and
soil organics
Phytoremediation
aromatics, chlorinated
aliphatics, hydrocarbons,
nutrients
rhizosphere microbial
degradation, plant-produced
enzymes
(Horner, 1995)
2 - 42
2.3.1 Sedimentation
Sedimentation is the removal of suspended particulates from the water column by gravitational
settling. This can be a major mechanism of pollutant removal, particularly soil particles and TSS
in BMPs such as ponds and constructed wetlands. The settling of discrete particles is dependent
upon particle size and settling velocity; fluid density; fluid viscosity; particle diameter and shape;
turbulence or short-circuiting; peak flowthrough rate; and volume of water (Stahre and Urbonas,
1990). Pollutants such as metals, hydrocarbons, nutrients and oxygen demanding substances can
become adsorbed or attached to particulate matter, particularly clay soils; removal of stormwater
particulates by sedimentation helps remove a large portion of pollutants associated with
particulates from the water column. Stormwater BMPs that utilize settling are usually suited for
dual purposes; that is, they can also provide storage volume for peak rate control, channel erosion
and/or flood control (VA DCR, 1999).
The main factor governing the efficiency of a BMP at removing suspended matter by
sedimentation is the time available for particles to undergo settling. Numerous research
investigations show that significant settling of urban pollutants occurs in the first 6 to 12 h of
detention, and fine particulates such as clay and silt require detention times of days or even
weeks to settle out of suspension (Clar et al., 2003). Evaluating the settling characteristics of the
particulates in runoff becomes a prerequisite to designing a BMP in order to determine the
detention time necessary for adequate settling to occur. Settling of particulates is also dependent
upon the initial concentration of suspended solids (SS) in the runoff; runoff with higher initial
concentrations of SS will have a greater removal efficiency.
The ability of pollutants to settle is directly influenced by particle size. Smaller particles take
longer to settle, and conversely, the larger the particle, the faster is its settling velocity. However,
particle size is not the only factor governing settling ability. The settling ability also depends on
the difference between the density of the fluid suspending the particle and the density of the
particle. Large, dense particles such as sand will fall through fluid at a faster rate than smaller,
less dense particles, such as clay. The volume of particles suspended within the fluid also
governs the settling rate. The more particles are suspended within the fluid, the faster the rate of
settling, but at some point the rate of settling will bottom out (Clar et al., 2003).
Turbulence, eddies, multilayered flows, circulation currents, and diffusion at inlets and outlets
affect the settling ability of particles as each of these factors can resuspend particles into the
water column. A decrease in flowthrough rate and surface loading improves sedimentation with
the most significant difference observed for larger particles. Actual field conditions must take
into account the particle settling velocity and surface loading rates during runoff conditions as
sediment removal under these conditions varies with storm intensity. The size of the body of
water relative to stormwater runoff will also determine the settling ability of sediment. The
larger the stormwater loading rate, the lower the removal of sediment by settling. Settling also
occurs after stormwater is trapped and ponded between storms. Because the intervals between
storm events occur randomly, understanding the effective ratio of storage volume to mean runoff
2 - 43
rate and the ratio of sediment volume removed to mean runoff rate is essential to predicting longterm averages (Clar et al., 2003).
The most widely used stormwater management practices that employ sedimentation are retention
and detention structures such as ponds and constructed wetlands. These can be designed to
effectively remove sediment from stormwater. Stormwater management basins with a permanent
pool of water have a removal percentage of total suspended solids of about 50 to 90%
(MWCOG, 1983; OWML, 1983; Dorman et al., 1989; City of Austin, 1990). Extended
detention ponds have a similar percentage of removal (MWCOG, 1983; City of Austin, 1988;
City of Austin, 1990). However, detention ponds may have a lower sediment removal efficiency
over the long-term than retention ponds, as seen from a few investigations (Clar et al., 2003).
This is because of the resuspension of sediments deposited on the detention pond bed from
previous storm events, during a new storm event.
Sedimentation removes particulates and other dissolved materials adsorbed to settleable particles
in a structural BMP, yet, the removal rate by settling of pollutants other than sediment particles is
inconclusive. One of the reasons for this confusion is a lack of understanding regarding the
specific process(es) in a BMP responsible for removing the pollutant(s) of concern. In retention
ponds, for example, processes such as settling, biological uptake, volatilization, infiltration to
groundwater, and adsorption occur simultaneously. While nitrogen, phosphorus, and bacteria
may be removed to some extent by adsorption to larger particulates, this is not expected to be a
primary mechanism for their treatment. Metals, however, are present in particulate and dissolved
form and some metal species are removed by coagulation and sedimentation.
2.3.2 Filtration
Filtration is typically limited to BMPs that address water quality and involves the removal of
particulates by passing stormwater through a porous medium to strain the pollutants out of
stormwater. Commonly used media in stormwater BMPs include soil, sand, gravel, peat,
compost and various combinations such as peat/sand, soil/sand and sand/gravel. Existing media
filtration practices commonly use trenches filled with sand or peat. Since the stormwater must
pass through the filter media in order to be treated, these structures are limited to small drainage
area (less than 5 acres) and low flow rates. A drawback to these structures is the overflow or
bypass of large flows from high intensity storms. Filtration is a complex process that depends on
a number of variables, including particle shape and size, void size in filter media, and velocity of
the fluid moving through the media. Solids and associated pollutants such as metals and
nutrients could be removed by filtration. Organic filtration media such as peat or leaf compost
are also effective in removing soluble nutrients from urban runoff (Metro Council, 2001).
Typically, stormwater filters remove particulates and adsorbed pollutants, such as sediment,
organic carbon, phosphorus, and many trace metals. Particulate pollutants are either trapped by
cation/anion exchange or are prevented from moving beyond the filter. However, in some cases,
the filtration process can increase the pollutant level of stormwater. The majority of N species
2 - 44
associated with particulate matter are organic nitrogen compounds. Filters that inadvertently
become anaerobic may cause nitrification of these species, and can release ammonia and nitrate
into stormwater (Clark et al., 2001). Once the treatment volume is achieved during a given
storm, the excess runoff bypasses the filter and remains untreated. Filtration of infiltrated flow to
remove sediments is more complicated, consisting of interception, straining, flocculation and
sedimentation as the water percolates through the granular subsurface (Cammermayer et al.,
2000). The process is augmented as clay particles are adsorbed to positively charged organic
matter, which can enhance settling in surface flows and improve retention in the subsurface.
However, in vegetated stormwater channels, the sorption mass rate is quite low relative to
sedimentation rates.
2.3.3 Sorption
The clay and organic matter in soil hold a negative charge. The ability of soil organic matter to
hold cations such as calcium and aluminum represent the soil's cation exchange capacity. This
process is most readily used to remove pollutants from stormwater. Organic matter such as peat
or leaf matter in the filter media uses its cation exchange capacity to bind pollutants to the filter.
The treatment of all runoff through filter media and biofilters, such as the bioretention cell (Clar
and Green, 1993) are other examples of cation exchange processes. The media traps particulates
(through filtration), adsorbs organic chemicals, and removes up to 90% of solids, 85% of oil and
grease, and 82 to 98% of heavy metals (through cation exchange from leaf decomposition).
While adsorption is not a common mechanism used in stormwater BMPs, it can still occur in
infiltration systems where the underlying soils contain appreciable amounts of clay; in organic
filters; or in wetland systems. Dissolved metals that are contained in stormwater runoff can be
bound to the clay particles as stormwater runoff percolates through clay soils in infiltration
systems.
The extent to which a given metal is adsorbed is affected by a number of factors, including
competitive effects of other ionic metals; the presence of iron and manganese oxides and organic
carbon; and, especially, pH (Clar et al., 2003). Treatment trains that include adsorptive media
may provide effective treatment for dissolved metals. Such media include compost, granulated
activated carbon, or diatomaceous earth, all of which work on a cation exchange principle.
2.3.4 Phytoremediation
Plants are able to degrade (break down) organic pollutants through their metabolic processes, and
the use of aquatic plants to treat wastewater, and the use of wetlands to treat farming effluent and
mining runoff is well known (Yu and Nawang, 1993). Phytoremediation is an umbrella term that
covers many different plant-based approaches for cleaning up contaminated environments, and
refers to the use of plants to degrade, sequester and stabilize organic and metal pollutants in
stormwater (U.S. EPA, 1998b). In simple terms, this means rendering pollutants harmless by
using green plants to remove them from the environment. More recently, the bacterial activity
associated with the roots of grasses and other plants has been explored for its organic degradation
2 - 45
potential (enhanced rhizosphere phytodegradation). The efficiency of phytoremediation may
vary depending on the depth of soil and the type and species of pollutants in water that are most
available for plant uptake (U.S. EPA, 2001b).
Phytoremediation is a recent technology with immense development potential. The positive
effects of plants can be both direct and indirect, and include:
<
increased microbial degradation in the rhizosphere, including co-metabolism,
<
uptake and accumulation in roots and foliage,
<
degradation in the plant,
<
volatilization of the compounds, and
<
plant-produced enzymes which degrade pollutants (Rasmussen and Olsen, 2004).
Phytoremediation can involve any of the following approaches. Phytoextraction uses the ability
of plants to take up and remove contaminants from soil and water and accumulate them in plant
tissues, which may then be harvested and removed from the site. The use of plants and (or) their
associated microbes to volatilize contaminants (volatile organic compounds, i.e., solvents, and
recently, inorganics such as Hg, Se, etc.) from soil or water is known as phytovolatilization. In
this process, plants take up water containing organic contaminants and release the contaminants
or the breakdown products into the air through their leaves. Se volatilization appears to be a
significant pathway of Se removal from contaminated sites (Terry, 2001b). Although
transferring contaminants to the atmosphere may not achieve the goal of complete remediation,
phytovolatilization may still be desirable in that it reduces prolonged soil exposure and the
associated risk of groundwater contamination; another advantage is that there is no hazardous
waste generation that warrants proper disposal measures as may be the case in phytoextraction.
Phytodetoxification involves the ability of plants to change the chemical species of the
contaminant to a less toxic form. Research shows the ability of plants to take up toxic Cr(VI)
species and convert it into the non toxic trivalent Cr(III). Phytostabilization uses plants to
immobilize contaminants chemically and physically at the site, thereby preventing their
movement to surrounding areas (Terry, 2001a).
Phytotransformation, also referred to as phytodegradation, is the breakdown of organic
contaminants sequestered by plants via (i) metabolic processes within the plant, or (ii) the effect
of compounds, such as the enzymes deoxygenase and halogenase, which are produced by the
plant. The organic contaminants are degraded into simpler compounds that are integrated with
plant tissue, which in turn, foster plant growth. Remediation of a site by phytotransformation is
dependent on the direct uptake of contaminants from the media and accumulation in the
vegetation. Certain enzymes produced by plants are able to breakdown and convert chlorinated
solvents (e.g., trichloroethylene), ammunition wastes, and herbicides. This technology can also
be used to remove contaminants from petrochemical sites and storage areas, fuel spills, landfill
leachates, and agricultural chemicals. Successful implementation of this technology requires that
the transformed compounds that accumulate within the plant be non toxic or significantly less
toxic than the parent compounds. Phytotransformation may also be used in concert with other
remediation technologies or as a polishing treatment. For example, a combination of
2 - 46
phytoremediation using orchard grass and a soil/sand filter material can efficiently treat creosotecontaminated groundwater (Rasmussen and Olsen, 2004).
The direct uptake of chemicals into plant tissue via the root system is dependent on uptake
efficiency, transpiration rate, and concentration of the chemical in soil water. Uptake efficiency
depends on chemical speciation, physical/chemical properties, and plant characteristics, whereas
transpiration rate depends on plant type, leaf area, nutrients, soil moisture, temperature, wind
conditions, and relative humidity. Two processes of remediation can occur after the organic
compound has been translocated by the plant: (i) storage of the chemical and its fragments into
the plant via lignification, and (ii) complete conversion to carbon dioxide and water. These
techniques have been successfully employed to treat and remove the following contaminants in
the environment: aromatics (BTEX); chlorinated aliphatics (TCE); herbicides (atrazine,
alachlor); hydrocarbons (TPH); nutrients (NO3 ­, NH4 +, PO4 3-) (U.S. EPA, 1998b).
2.3.5 Biological Processes
Biological processes are one of the most effective types of removal mechanisms for soluble
pollutants such as nutrients in stormwater runoff and an important mechanism of nutrient control
in stormwater BMPs (VA DCR, 1999). Vegetated lands help prevent erosion and remove
contaminants from surface runoff in a process known as bioassimilation. It also helps to slow
down the runoff, giving it more time to infiltrate into the soil. A combination of shallow
permanent pool depths and abundant vegetation help to create conditions that allow a natural
food chain to develop; marsh plants, algae and microorganisms that grow on the shallow organic
rich sediments take up soluble forms of nutrients needed for their growth. Periodic harvesting of
this vegetation allows for permanent removal of these nutrients. BMPs suited for this pollutant
removal mechanism include enhanced ED as well as retention ponds, constructed stormwater
wetlands, and in some cases bioretention. Some wetland plants not only assimilate
contaminants; wetland plants such as cattails have been shown to assimilate, and in some cases,
transform pathogenic bacteria and metals (1994). The use of submerged aquatic vegetation as
the dominant vegetation in treatment wetlands in south Florida results in a higher P removal
performance than wetlands dominated by rooted, emergent plants, with the newly accreted
sediment being the likely, long-term sink for this P (Knight et al., 2003). Macrophytes reduce
flow velocity, increase sedimentation and retention by deposition in running waters and
contribute to total monthly phosphorus retention (up to 25%) by increasing deposition of
particulate organic matter (Schulz et al., 2003).
Many stormwater BMPs utilize a combination of these pollutant removal mechanisms. In some
cases, development of an organic layer occurs within a BMP that has been in operation for a
period of time, thereby increasing the adsorption potential of the BMP. BMPs that include plants
and grasses also display increased pollutant removal efficiency over time as the biomass
increases. This increase in biomass slows the velocity of the runoff through the BMP and allows
for increased gravitational settling and filtering of pollutants, as well as decreased export of
sediment and attached pollutants via erosion (VA DCR, 1999).
2 - 47
2.4 PERFORMANCE OF STRUCTURAL BMPS
The pollutant removal capability of a BMP is primarily governed by three inter-related factors:
the removal mechanisms used; the fraction of the annual runoff volume that is effectively treated;
and, the nature of the urban pollutant being removed. The nature of the pollutant being removed
often sets an upper limit on the potential removal rate that can be achieved. While particulate
pollutants such as sediment and lead are relatively easier to remove by common BMP removal
mechanisms including settling and filtering, it is not the case with soluble pollutants such as
nitrate, phosphates and some trace metals. The above-mentioned removal mechanisms have little
or no effect and require the additional use of biological methods such as uptake by bacteria,
algae, rooted aquatic plants or terrestrial vegetation. Most BMPs can achieve an extremely high
removal rate for suspended sediment and trace metals that exist largely in particulate forms.
However, removal rates that are much lower are generally obtained for total phosphorus,
oxygen-demanding materials, and total nitrogen, since they typically exist in both particulate and
soluble forms. Tables 2-4 summarizes the processes governing pollutant removal in commonly
used structural BMPs.
Table 2-4. Processes Governing Pollutant Removal in Commonly Used Structural BMPs
Pollutant
Constituents
Structural BMP Type and Process Mechanisms
Pond
Wetland
Biofilters
Infiltration
Sand Filter
Heavy metals
Sorption
Settling
Sorption
Settling
Phyto­
remediation
Sorption
Filtration
Sorption
Filtration
Phyto­
remediation
Settling
Sorption
Filtration
Toxic Organics
Sorption
Bio­
degradation
Settling
Phyto­
volatilization
Sorption
Bio­
degradation
Settling
Phyto­
volatilization
Sorption
Filtration
Sorption
Filtration
Settling
Phyto­
volatilization
Sorption
Filtration
Nutrients
Bioassimilation
Bioassimilation
Phyto­
remediation
Sorption
Sorption
Bioassimilation
Phyto­
remediation
Sorption
2 - 48
Solids
Settling
Sorption
Settling
Sorption
Filtration
Sorption
Filtration
Settling
Filtration
Oil & Grease
Sorption
Settling
Sorption
Settling
Sorption
Sorption
Settling
Sorption
BOD5
Bio­
degradation
Bio­
degradation
Bio­
degradatio
n
Bio­
degradation
Bio­
degradation
Filtration
Filtration
Settling
Filtration
Predation
UV
irradiation*
(sunlight)
Sedimentation
Aggregation
Oxidation
Antibiosis
(Scholze et al., 1993) *(Davies and Bavor, 2000)
Pathogens
Settling
UV
irradiation
The efficiency of a given BMP in removing pollutants is dependent upon a number of
site-specific variables: (i) size, type and design of the BMP; (ii) soil type and characteristics; (iii)
geology and topography of the site; (iv) intensity and duration of rainfall; (v) length of antecedent
dry periods; (vi) climatological factors such as temperature, solar radiation, and wind; (vii) size
and characteristics of the contributing watershed; and, (viii) properties and characteristics of the
various pollutants (U.S. EPA, 1999b). Urban stormwater BMP performance is affected by
various factors as shown in Table 2-5.
Table 2-5. Factors affecting the Performance of Structural BMPs
Design Characteristics
Detention time
Storage volume
Shape of BMP (length : width ratio of
detention pond, slope, flow path
characteristics, buffer width)
Pond surface area relative to contributing
catchment area
Type of substrate/sediment
Auxiliary devices (e.g., baffles, sediment
forebay)
Presence of vegetation
2 - 49
Processes
Sedimentation
Filtration
Adsorption
Volatilization
Oxidation
Precipitation
Biological nitrification
Microbial decomposition
Phytoremediation
Environmental Conditions
Storm intensity, loading rate
Drainage basin land cover
Physical-chemical properties of BMP surface
water
Particle size distribution and settling velocity
Pre-storm water quality in BMP
Thermal stratification
(Law and Band, 1998)
2.4.1 Suspended Solids, Nutrients, and Heavy Metals, and BMP Performance
BMPs are primarily designed to remove TSS and pollutants sorbed to particles, with gravitational
settling being the predominant process for pollutant removal (Schueler et al., 1992). The ability
to treat TSS is a function of particle size distribution, storm intensity, loading rate, and geometry
and age of BMP facility. An increase in TSS concentration observed with increased storm
intensity could be due to larger size fractions of TSS, that result in a greater removal efficiency
(Ferrara and Witkowski, 1983). Pond geometry also influences the sedimentation rate in that the
finer particles do not settle out when the length-to-width ratio is insufficient; in detention ponds
draining a commercial complex, observations of sediment sorting showed coarsest particles
settling nearest the inlet (Marsalek et al., 1997). The effectiveness of sedimentation is further
limited by the available storage volume and hydraulic loading rate of runoff to the BMP. BMP
performance is adversely affected under the following conditions:
<
if the runoff volume is greater than the storage volume of the BMP,
<
short-circuiting within the pond if sediment is allowed to accumulate and consequently
reduce storage volume (U.S. EPA, 1993b), and
<
age of the BMP.
In order to improve stormwater quality when the water is polluted with excessive nutrient levels,
BMP facilities are designed to provide sedimentation for particle-phase nutrients and biological
uptake for soluble nutrients (Martin, 1988). Wet detention ponds and constructed wetlands act as
a sink, source or transformer of nutrients. However, the performance of BMPs for
nutrient-enriched stormwater is unpredictable. This could be due to the complex nature of
nutrient speciation, which is affected by seasonality, detention time, organic matter content,
2 - 50
oxygen availability and plant biomass. The most effective phosphorus removal mechanisms are
adsorption, complexation, precipitation reactions with Al, Fe, Ca and clay particles, and by peat
accretion. Pond conditions amenable to nitrification-denitrification processes are most effective
for nitrogen removal. Similar to TSS, removal efficiencies of nutrients in BMPs are affected by
particle size, quality of the substrate and age of the treatment facility. BMP performance is
highly variable for nutrients compared to other constituents with regard to both particulate and
dissolved phases, as well as different species of nutrients (Tanner et al., 1997). Also, the soluble
phase of nutrients may not be effectively treated in a wet detention basin. Studies on wet
detention facilities show no clear relationship between particle size distribution and storm
intensity to explain the variability of reduction for all particle size classes for total Kjeldahl
nitrogen (TKN) and total phosphorus (TP). The net export of pollutants may be attributed to the
large proportion of pollutants associated with the soluble size fraction of solids (i.e., < 1µm)
(Law and Band, 1998).
BMPs generally contribute to the reduction of total heavy metal pollutants from urban
stormwater. Overall, studies report a moderate to high removal efficiency for wet detention
basins, constructed wetlands, and combined pond-wetland systems (Schueler et al., 1992). It is
assumed that pollutants in urban stormwater are largely sorbed to suspended sediments and
therefore sedimentation is considered to be the predominant removal mechanism for heavy
metals. The concentration of heavy metals in sediments indicates the improvement that BMPs
have on stormwater quality (Marsalek et al., 1997). Despite the strong ability of particulates and
organic matter to sequester heavy metals, TSS are not good surrogates for other pollutants (Law
and Band, 1998). This however, counters a design principle for BMPs which assumes that an
80% removal of TSS will control other pollutants, including heavy metals (U.S. EPA, 1993b).
Illustrating the presence of heavy metals in urban runoff in the dissolved phase implies a low
reliance on sedimentation to remove heavy metals from the water column. Adverse effects
induced by sediment contamination may counter the benefits derived from improved water
quality. The solubility of heavy metals is affected by cation-exchange capacity, sensitivity to pH
and Eh, which in turn affects the ability of sediments to sequester heavy metals from the water
column. Heavy metal mobilization from sediments can be enhanced due to intermittent flooding,
which produces alternating periods of aerobic and anaerobic conditions. Also, the change in the
oxidized state of an environment can alter the speciation of metals, and hence their
bioavailability.
REFERENCES
Andoh, R.Y.G. and A.J. Saul. (2002). The Use of Hydrodynamic Vortex Separators and
Screening Systems to Improve Water Quality. In: 3rd International Conference on Sewer
Processes and Networks Proceedings, Paris, France.
Atlanta Regional Commission. (2001). Georgia Stormwater Management Manual Volume 2:
Technical Handbook. Atlanta Regional Commission, www.georgiastormwater.com
Bradner, J.N. and M.P. Wanielista. (1992). Project SMART. Florida Department of
2 - 51
Environmental Regulation, Orlando, FL.
State of California Department of Transportation (Caltrans). (2002). Proposed Final Report:
BMP Retrofit Pilot Program. CTSW-RT-01-050. California Department of
Transportation Report CTSW-RT-01-050 , and Sacramento, CA.
Cammermayer, J.W., R.R. Horner, and N. Chechowitz. (2000). Vegetated Stormwater Facility
Maintenance. Research Report. Prepared for Washington State Transportation
Commission, Department of Transportation (WSDOT) in cooperation with U. S.
Department of Transportation, Federal Highway Administration (FHWA), Seattle, WA.
California Stormwater Quality Association (CASQA). (2003). Stormwater Best Management
Practice Handbook: New Development and Redevelopment. www.cabmphandbooks.com.
City of Austin. (1988). Environmental Criteria Manual. Environmental Resource Management
Division, Austin, TX.
City of Austin. (1990). Removal Efficiencies of Stormwater Control Structures. Final Report.
Austin, TX.
Clar, M., B.J. Barfield, and S. Yu. (2003). Considerations in the design of treatment best
management practices (BMPs) to improve water quality. EPA 600/R-03/103. U.S.
Environmental Protection Agency, Office of Research and Development, Cincinnati, OH.
www.epa.gov/ORD/NRMRL/Pubs/600r03103/600r03103.htm October 2003.
Clar, M.L. and R. Green. (1993). Design Manual for Use of Bioretention in Stormwater
Management. Department of Environmental Resources Watershed Protection Branch,
Prince George's County, MD.
Clark, S., R. Pitt, and P. Brown. (2001). Effect of Anaerobiosis on Filter Media Pollutant
Retention. In: Information and Monitoring Needs for Evaluating the Mitigating Effects of
BMPs - Engineering Foundation and the American Society of Civil Engineers
Conference, Snowmass, CO.
Center for Watershed Protection (CWP). (1998). Better Site Design: A Handbook for Changing
Development Rules in Your Community. Ellicott City, MD.
CWP. (2000). An Introduction to Better Site Design. In: The Practice of Watershed Protection,
Schueler, T.R. and H.K. Holland, (editors). Center for Watershed Protection,, Ellicott
City, MD. Pages 623-632.
Davies, C.M. and H.J. Bavor. (2000). The fate of stormwater-associatied bacteria in constructed
wetland and water pollution control pond systems. Journal of Applied Microbiology 89:
349-360.
D'Itri, F.M., (editor). (1992). Chemical Deicers and the Environment. Lewis Publishers, Inc.,
Chelsea, MI.
Dorman, M.E., J. Hartigan, R.F. Steg, and T. Quasebarth. (1989). Retention, detention and
overland flow for pollutant removal from highway stormwater runoff. Volume I. Report
No. FHWA/RD 89/202. Federal Highway Administration, Washington, D. C.
Ferrara, R. and P. Witkowski. (1983). Stormwater quality characteristics in detention basins.
Journal of Environmental Engineering 190(2): 428-447.
FHWA. (2000). Stormwater Best Management Practices in an Ultra-Urban Setting: Selection
and Monitoring. U.S. Department of Transportation, Washington, D. C.
http://www.fhwa.dot.gov/environment/ultraurb/index.htm
2 - 52
Field, R. and T.P. O'Connor. (1996). Swirl Technology: Enhancement of Design, Evaluation and
Application. Journal of Environmental Engineering 122(8): 741-748.
Field, R., D. Averill, T.P. O'Connor, and P. Steel. (1997). Vortex Separation Technology. Water
Quality Research Journal of Canada 32(1): 185-214.
Field, R. and D. Sullivan, (editors). (2003). Wet-Weather Flow in the Urban Watershed:
Technology and Management. CRC Press LLC, Boca Raton, FL.
Florida DER. (1988). The Florida Development Manual: A Guide to Sound Land and Water
Management. Florida Department of Environmental Regulation, Tallahassee, FL.
GreenRoofs. Exploring the Ecology of Organic Green Roof Architecture, www.greenroofs.com.
Greater Vancouver Sewerage and Drainage District (GVSDD). (1999). Best Management
Practices Guide for Stormwater. Burnaby, B. C.
Horner, R. (1995). Constructed Wetlands for Urban Runoff Water Quality Control. Department
of Civil Engineering, University of Washington, Seattle, WA.
Knight, L.R., B. Gu, A.R. Clark, and M.J. Newman. (2003). Long-term phosphorus removal in
Florida aquatic systems dominated by submerged aquatic vegetation. Ecological
Engineering 20: 45-63.
Konieek, Z., K. Pryl, and M. Suchanek. (1996). Practical Applications of Vortex Flow Separators
in the Czech Republic. Water Science and Technology 33(9): 253-260.
Law, N.L. and L.E. Band. (1998). Performance of urban stormwater best management practices,
http://www.unc.edu/depts/geog/them/projects/BMP.html. University of Carolina - Chapel
Hill, Department of Geology, Chapel Hill, NC.
Lee, J., K. Bang, J. Choi, L.H. Ketchum Jr, and Y. Cho. (2003). The vortex concentrator for
suspended solids treatment. Water Science and Technology 47(9): 335-341.
Low Impact Development Center (LIDC). Green Roofs. Beltsville, MD. http://www.lidstormwater.net/greenroofs/greenroofs_home.htm.
Liptan, T. and E. Strecker. (2003). EcoRoofs (Green Roofs) - A More Sustainable Infrastructure.
In: National Conference on Urban Stormwater: Enhancing Programs at the Local Level.
Proceedings. EPA-625-R-03-003, February 17-20, Chicago, IL.
http://www.epa.gov/owow/nps/natlstormwater03/20Liptan.pdf.
Marsalek, J., W.E. Watt, B.C. Anderson, and C. Jaskot. (1997). Physical and chemical
characteristics of sediment from a stormwater management pond. Water Quality Research
Journal of Canada 32(1): 89-100.
Martin, E.H. (1988). Effectiveness of an urban runoff detention pond - wetland system. Journal
of Environmental Engineering 114(4): 810-827.
Maryland Department of the Environment (MDE). (2000). Maryland Stormwater Design Manual
Volumes I & II. Center for Watershed Protection and Maryland Department of the
Environment, Baltimore, MD. http://www.mde.state.md.us.
Michigan Department of Environmental Quality (MDEQ). (1998). Guidebook of Best
Management Practices for Michigan Watersheds. Surface Water Quality Division.
(Metro Council). (2001). Minnesota Urban Small Sites BMP Manual. Stormwater Best
Management Practices for Cold Climates. Prepared for the Metropolitan Council by Barr
Engineering Company,, St. Paul, MN.
Metropolitan Washington Council of Governments (MWCOG). (1983). Urban Runoff in the
2 - 53
Washington Metropolitan Area - Final Report Washington D. C. Area Urban Runoff
Project. Prepared for U.S. EPA and WRPB, Washington, DC.
MWLAP. (1992). Best Management Practices to Protect Water Quality,
http://wlapwww.gov.bc.ca/wat/wq/nps/BMP_Compendium/BMP_Introduction/bmphome
.htm. Ministry of Water, Land and Air Protection, Government of British Columbia,.
Northern Virginia Planning District Commission (NVPDC). (1979). Guidebook for Screening
Urban Nonpoint Pollution Management Strategies. Annandale, VA.
Northern Virginia Planning District Commission (NVPDC). (1996). NonStructural Urban BMP
Handbook: A guide to nonpoint source pollution prevention and control through
nonstructural measures. Annandale, VA.
Northern Virginia Planning District Commission and Engineers and Surveyors Institute
(NVPDC and ESI). (1992). Northern Virginia BMP Handbook: A guide to planning and
designing best management practices in northern Virginia. Annandale, VA.
New York State Department of Environmental Conservation (NYSDEC). (2001). New York state
stormwater management design manual (Draft). Division of Water, Albany, NY.
http://www.dec.state.ny.us/website/dow/swmanual. October 2001.
Occoquan Watershed Monitoring Laboratory (OWML). (1983). Final Contract Report:
Washington Area NURP Project. Prepared for Metropolitan Washington Council of
Governments, Manassas, VA.
Prince George's County Department of Environmental Resources (PGDER). (1993). Design
Manual for Use of Bioretention in Stormwater Management. Division of Environmental
Management, Watershed Protection Branch, Landover, MD.
Pierce County Public Works and Utilities (Pierce County). Stormwater Pollution Prevention
Manual - A Guide to Best Management Practices for Industries, Businesses and
Homeowners. University Place, WA.
http://www.co.pierce.wa.us/pc/services/home/environ/water/swm/sppman/bmps9.htm.
Pitt, R., B. Robertson, P. Barron, A. Ayyoubi, and S. Clark. (1999). Stormwater Treatment at
Critical Areas: The Multi-Chambered Treatment Train (MCTT). EPA-600-R-99-017.
National Risk Management Research Laboratory, United States Environmental Protection
Agency, Cincinnati, OH.
Prince George's County (DER). (2002). Low Impact Development Integrated Management
Practices Guide Book. government in Prince George's County, Maryland,
Rasmussen, G. and R.A. Olsen. (2004). Sorption and biological removal of creosotecontaminants from groundwater in soil/sand vegetated with orchard grass (Dactylis
glomerata). Advances in Environmental Research 8: 313-327.
Scholze, R., V. Novotny, and R. Schonter. (1993). Efficiency of best management practices for
controlling priority pollutants in runoff. Water Science and Technology 28(3-5): 215-224.
Schueler, T.R. (1987). Controlling Urban Runoff: A practical manual for planning and
designing urban BMPs. Metropolitan Washington Council of Governments, Washington,
DC.
Schueler, T.R. (1992). Design of Stormwater Pond Systems. Metropolitan Washington Council
of Governments,, Washington, D. C.
Schueler, T.R., P.A. Kumble, and M.A. Hearty. (1992). A Current Assessment of Urban Best
2 - 54
Management Practices: Techniques for reducing nonpoint source pollution in the coastal
zone. Metropolitan Washington Council of Governments, Washington, DC.
Schulz, M., H.-P. Kozerski, T. Pluntke, and K. Rinke. (2003). The influence of macrophytes on
sedimentation and nutrient retention in the lower river Spree (Germany). Water Research
37: 569-578.
Schwarz, T. and S. Wells. (1999). Stormwater Particle Removal using Cross-Flow Filtration and
Sedimentation. In: Advances in Filtration and Separation Technology Volumes 13 A & B,
Leung, W., Ptak, T.,, (editor). American Filtrations and Separations Society. Pages 219­
226.
Stahre, P. and B.R. Urbonas. (1990). Stormwater Detention for Drainage, Water Quality and
CSO Control. Prentice Hall, Englewood Cliffs, NJ.
Strecker, E. (2001). Low-Impact Development (LID) Is it really low or just lower? In: Linking
Stormwater BMP Designs and Performance to Receiving Water Impact Mitigation ­
Proceedings of an Engineering Foundation Conference, Urbonas, B.R., (editor).
American Society of Civil Engineers, Snowmass Village, CO. Pages 210-222.
Tanner, C.C., J.P.S. Sukias, and M. Upsdell. (1997). Relationships between loading rates and
pollutant removal during maturation of gravel-bed constructed wetlands. Journal of
Environmental Quality 27: 448-458.
Taylor, A. and T. Wong. (2002). Nonstructural Stormwater Quality Best Management Practices
- An Overview of Their Use, Cost and Evaluation. Technical Report 02/11. Cooperative
Research Center for Catchment Hydrology, Victoria, NZ.
Terry, N. (2001a). http://www.hawaii.edu/abrp/Technologies/phytran.html).
Terry, N. (2001b). Use of flow-through constructed wetlands for the removal of selenium in
agricultural-tile drainage water. Annual Report 2000-2001. U.C Salinity/Drainage Task
Force, Division of Agriculture and Natural Resources, University of California, Berkeley,
CA.
United States Environmental Protection Agency (U.S. EPA). (1977). Swirl Device for Regulating
and Treating Combined Sewer Overflows. EPA Technology Transfer Capsule Report.
EPA-625-2-77-012. Washington, DC.
United States Environmental Protection Agency (U.S. EPA). (1983). Results of the Nationwide
Urban Runoff Program. Volume I - Final Report. EPA-832-R-83-112. Washington, DC.
United States Environmental Protection Agency (U.S. EPA). (1992). Stormwater Management
for Construction Activities: Developing Pollution Prevention Plans and Best
Management Practices. EPA-832-R-92-005. Washington, DC.
United States Environmental Protection Agency (U.S. EPA). (1993a). Guidance Specifying
Management Measures for Sources of Nonpoint Pollution in Coastal Waters. EPA 840-B92-002. . Washington, DC.
United States Environmental Protection Agency (U.S. EPA). (1993b). Handbook - Urban Runoff
Pollution Prevention and Control Planning. EPA-625-R-93-004. Washington, DC.
United States Environmental Protection Agency (U.S. EPA). (1993c). Investigation of
Inappropriate Pollutant Entries into Storm Drainage Systems EPA-600-R-92-238.
Washington, DC.
United States Environmental Protection Agency Region 5 (U.S. EPA). (1998a). Illegal Dumping
2 - 55
Prevention Guidebook. Chicago, IL.
United States Environmental Protection Agency (U.S. EPA). (1998b). A Citizen's Guide to
Phytoremediation. EPA-542-F-98-011. Office of Solid Waste and Emergency Response,
Washington, DC.
United States Environmental Protection Agency (U.S. EPA). (1999a). Stormwater Management
Fact Sheet. EPA-832-F-99-001 through EPA-843-F-99-050. Washington, DC.
United States Environmental Protection Agency (U.S. EPA). (1999b). Preliminary Data
Summary of Urban Stormwater Best Management Practices. EPA-821-R-99-012. Office
of Water, Washington, DC.
United States Environmental Protection Agency (U.S. EPA). (2000a). Low Impact Development
Design Strategies: An integrated design approach. Prepared by Tetra Tech Inc., Fairfax,
VA for Department of Environmental Resources, Prince George's County MD, funding
provided by the United States Environmental Protection Agency, Washington, DC.
United States Environmental Protection Agency (U.S. EPA). (2000b). Low Impact Development
Hydrologic Analysis. Prepared by Tetra Tech Inc., Fairfax, VA for Department of
Environmental Resources, Prince George's County MD, funding provided by the United
States Environmental Protection Agency, Washington, DC.
United States Environmental Protection Agency (U.S. EPA). (2000c). Vegetated Roof Cover,
Philadelphia, Pennsylvania. EPA-841-B-00-005D. Washington, DC.
United States Environmental Protection Agency (U.S. EPA). (2000d). Low Impact Development
(LID) - A Literature Review. EPA-841-B-00-005. Washington, DC.
United States Environmental Protection Agency (U.S. EPA). (2001a). National Menu of Best
Management Practices for Stormwater Phase II. Washington, DC.
http://cfpub.epa.gov/npdes/stormwater/menuofbmps/menu.cfm.
United States Environmental Protection Agency (U.S. EPA). (2001b). A Citizen's Guide to
Phytoremediation. EPA-542-F-01-002. Washington, DC.
http://www.epa.gov/superfund/sites.
United States Environmental Protection Agency (U.S. EPA). (2003). National Conference on
Urban Stormwater: Enhancing Program at the Local Level Proceedings Chicago, IL
February 17-20, 2003. EPA-625-R-03-003. National Risk Management Laboratory,
Office of Research and Development, Cincinnati, OH.
UNEP IETC. (2000). Environmentally Sound Technologies in waste water treatment for the
implementation of the UNEP Global Programme of Action (GPA) "Guidance on
Municipal Wastewater" in collaboration with Murdoch University Environmental
Technology Centre,
http://www.unep.or.jp/ietc/publications/freshwater/sb_summary/index.asp. United
Nations Environment Programme International Environmental Technology Centre.
University of Rhode Island. (1994). Technical Analysis Report. Prepared for The Saugatucket
River Heritage Corridor Coalition by the Advanced Planning Studio of the Graduate
Curriculum in Community Planning and Area Development,
http://saugatucket.org/docs/SaugTechRpt/Saugatucket_Tech_Report.pdf.
Commonwealth of Virginia Department of Conservation and Recreation (VA DCR). (1999).
Virginia Stormwater Management Handbook, Volume I,. Division of Soil and Water
2 - 56
Conservation., Richmond, VA.
Washington State Department of Ecology (WA DOE). (1991). Stormwater Management Manual
for the Puget Sound Basin - Public Review Draft. Olympia, WA.
Washington State Department of Ecology (WA DOE). (2001). Stormwater Management Manual
for Western Washington. Volume V - Runoff Treatment BMPs. Water Quality Program,
http://www.ecy.wa.gov/programs/wq/stormwater/index.html.
Walker, T.A. and T.H.F. Wong. (1999). Effectiveness of Street Sweeping for Stormwater
Pollution Control, Technical Report 99/08. Cooperative Research Center for Catchment
Hydrology, Melbourne.
Wanielista, M.P. and J.N. Bradner. (1992). Project SMART - A Restoration Demonstration.
University of Central Florida, Orlando, FL.
Water Environment Federation and American Society of Civil Engineers (WEF and ASCE).
(1998). Urban runoff quality management. 1-57278-039-8 and 0-7844-0174-8. Water
Environment Federation, Alexandria, VA.
Wong, T.H.F., R.M. Wootton, and D. Fabian. (1996). A Solid Separator Using a Continuous
Deflective System. In: Proceedings 7th International Conference on Urban Stormwater
Drainage, Hannover, Germany. IAHR/IAWQ Joint Committee on Urban Storm
Drainage.
Yu, S.L. and W.M. Nawang. (1993). Best Management Practices for Urban Stormwater Runoff
Control. In: Integrated Stormwater Management, Field, R., M.L. O'Shea, and K.K. Chin,
(editors). Lewis Publishers, Boca Raton, FL. Pages 191-206.H
2 - 57
3
Structural BMP Design Practices
Swarna Muthukrishnan, Richard Field and Daniel Sullivan
CONTENTS
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 2
3.2 BMP Design Criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 3
3.2.1 Unified Stormwater Sizing Criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 3
Water Quality Volume (WQv) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 4
Recharge Volume (Rev) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 5
Channel Protection Volume (Cpv)
............................... 3-5
3.2.2 BMP Performance Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 6
3.3 Design of Structural BMPs to Improve Water Quality . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 6
3.3.1 Dry Extended-detention (ED) Ponds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 7
3.3.1.1 Stormwater Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 8
3.3.1.2 Pollutant Removal Capability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 8
3.3.1.3 Design Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 9
3.3.2 Wet Ponds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 21
3.3.2.1 Stormwater Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 21
3.3.2.2 Pollutant Removal Capability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 21
3.3.2.3 Design Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 23
3.3.3 Stormwater Wetlands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 31
3.3.3.1 Stormwater Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 31
3.3.3.2 Pollutant Removal Capability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 31
3.3.3.3 Design Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 32
3.3.4 Grassed Swales . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 40
3.3.4.1 Stormwater Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 40
3.3.4.2 Pollutant Removal Capability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 41
3.3.4.3 Design Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 42
3.3.5 Vegetated Filter Strips (VFS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 51
3.3.5.1 Stormwater Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 51
3.3.5.2 Pollutant Removal Capability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 51
3.3.5.3 Design Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 52
3.3.6 Infiltration Trenches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 58
3.3.6.1 Stormwater Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 58
3.3.6.2 Pollutant Removal Capability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 59
3.3.6.3 Design Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 60
3.3.7 Porous Pavement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 71
3-1
3.3.7.1 Stormwater Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3.7.2 Pollutant Removal Capability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3.7.3 Design Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3.8 Sand and Organic Filters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3.8.1 Stormwater Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3.8.2 Pollutant Removal Capability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3.8.3 Design Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3 - 73
3 - 73
3 - 73
3 - 79
3 - 79
3 - 80
3 - 80
3 - 92
3.1 INTRODUCTION
The selection and successful design of structural BMPs for stormwater quality enhancement is
the cornerstone of stormwater management in newly developing and redeveloping urban areas.
The cost effectiveness of each control has to be considered and measured against the actual
environmental benefits realized. Design objectives for BMPs can be stated in terms of
technology (e.g., by specifying a particular control device) or in terms of quantitative effect (e.g.,
by specifying a required degree of control or a maximum allowable effect). Quantitative
objectives can be defined for both hydrological parameters and constituent removal performance
parameters. Some examples of objectives based on hydrological parameters include peak flow
rate and retention of a defined water volume for a specific period of time, while objectives based
on chemical parameters include percent removal of specific chemical constituents and effluent
concentration or mass discharge targets (WEF and ASCE, 1998).
In 1990, the U.S. EPA promulgated phase I of the NPDES (National Pollutant Discharge
Elimination System) program regulations for stormwater discharges from municipal separate
storm sewer systems (MS4s) (U.S. EPA, 1996), which required municipalities to reduce
pollutants in urban runoff to the maximum extent practicable (MEP). The definition of MEP for
the control of stormwater pollutant discharges has focused primarily on the application of
economically achievable management practices. Stormwater runoff rates and volumes vary
highly between storms; hence, the statistical probabilities of runoff events and their management
have to be considered in developing practices to meet the MEP goal. It is therefore imperative to
examine the hydrology of urban runoff and the type and size of storm runoff events to result in a
robust design that has a high probability of meeting these regulations (WEF and ASCE, 1998).
Stormwater management programs were developed by various states in the 1980s when the
prevailing outlook was that the quality of receiving streams could be sustained by controlling the
flooding caused by increases in runoff volume from new development. The objective of these
stormwater management facilities therefore was to control peak flows. Efforts on stormwater
management during this time addressed BMP design for flood control based on hydrological
procedures with little or no satisfactory guidelines or criteria set forth for water quality
management. Most of these designs were flow-based, with an emphasis on the reduction of postconstruction flows of the 2- and 10-yr storm events to pre-development levels in new
3-2
development; this peak management approach addressed only stream channel erosion concerns
besides providing adequate flood control in receiving waters and not water quality. However,
changes in land use patterns due to increasing urbanization has resulted in a large increase in
impervious cover and runoff, strongly highlighting the need to control both the quality and
quantity of runoff in order to prevent stream channel erosion. In response to the provisions of the
Clean Water Act, (originally enacted in 1977 with subsequent revisions), a number of activities
such as the Nationwide Urban Runoff Program (NURP) (U.S. EPA, 1983) were initiated to
characterize and quantify the water quality impacts of WWF, and municipalities began adopting
BMPs for pollutant removal.
A growing national awareness of the wide range of environmental impacts of runoff and
urbanization has resulted in BMPs being designed to control larger storms as well as smaller, but
more frequent storms to achieve additional ecological benefits that include stream channel
protection and restoration, groundwater infiltration, protection of riparian habitat and biota, and
minimized thermal impacts. Collected runoff has also been used for irrigation, toilet flushing,
and other non-potable purposes, including ponds and wetlands that also enhance urban aesthetics.
This redirected approach in considering a watershed in its entirety and using BMPs for water
quality improvement has led to several procedures being established to achieve removal of
pollutants from storm runoff. A few approaches include the mandatory requirement to remove
80% of total suspended solids (TSS) from new development in coastal zone states (Coastal Zone
Management Act (CZMA), 1972), and controlling the first flush of pollutants associated with a
storm, mandating the capture of the first 0.5 to 1 in. of runoff (typically generated in the first
hours of the 1-yr storm).
3.2 BMP DESIGN CRITERIA
3.2.1 Unified Stormwater Sizing Criteria
The objective of any stormwater design criteria is to protect receiving waters from adverse
impacts associated with urban runoff. This goal can be successfully accomplished by adopting a
unified approach to sizing stormwater BMPs, which is influenced by several factors. A few
examples include local hydrological conditions, rainfall-runoff pattern, the type of BMP to be
installed, the volume of stormwater that would be treated, degree of imperviousness, prevailing
stormwater regulations to be adhered to etc. As an example, the guidelines proposed by the
Maryland Department of Environment, which consists of five main quality characteristics of
stormwater, is presented in Table 3-1 (MDE, 2000) and briefly explained below.
3-3
Table 3-1. Summary Example of Unified Stormwater Sizing Criteria
Sizing Criteria
Description
Water Quality Volume
(WQv) (acre-ft)
WQv = [(P)(Rv)(A)]/12
P = 1.0 in. in Eastern Zone and 0.9 in. in Western Zone
Rv = volumetric runoff coefficient
A = Area in acres
Recharge Volume
(Rev) (acre-ft)
Rev = [(S)(Rv)(A)]/12
S = Soil Specific Recharge Factor
Rev is a sub-volume of WQv
Channel Protection
Storage Volume
(Cpv)
Cpv = 24 h (12 h in USE III and IV watersheds) extended-detention
of the post-developed 1-yr 24 h storm event. Not required for
direct discharges to tidal waters and the Eastern Shore of Maryland.
Overbank Flood
Protection Volume
(Qpx)
Local review authorities may require that the peak discharge from
the 10-yr storm event be controlled to the pre-development rate
(Qp10). No control of the 2-yr storm event (Qp2) is required.
Extreme Flood Volume
(Qf)
Consult with the appropriate local reviewing authority. Normally
no control is needed if development is excluded from the 100-yr
flood plain and downstream conveyance is adequate.
(MDE, 2000)
Water Quality Volume (WQv)
Water quality volume is the storage needed to capture and treat the runoff from 90% of average
annual rainfall. Numerically this is equivalent to an inch of rainfall multiplied by the volumetric
runoff coefficient (Rv) and site area. Treatment of the WQv shall be provided at all developments
where stormwater management is required. According to (MDE, 2000), a minimum WQv of 0.2
in./acre shall be met at sites or drainage areas that have less than 15% impervious cover, while
drainage areas having no impervious cover and no proposed disturbance during development
may be excluded from the WQv calculations. While the WQv is the storage volume needed to
capture and treat the runoff from 90% of the average annual rainfall, it also provides management
at a critical level (one-third bankfull elevation) within stream channels
Recharge Volume (Rev)
The criteria for maintaining recharge is based on the average annual recharge rate of the
hydrologic soil group(s) present at a site as determined by the United States Department of
Agriculture Natural Resources Conservation Service (USDA-NRCS) soil surveys or from
detailed soil investigations. Calculation of the specific recharge factor (S) for each soil is based
on the USDA-NRCS average annual recharge volume per soil type divided by the annual rainfall
(42 in. in the case of Maryland) and multiplied by 90%, consistent with the WQv methodology.
3-4
The recharge volume is considered part of the total WQv that must be addressed at a site and can
be achieved either by nonstructural techniques (e.g., buffers, runoff disconnection), structural
practices (e.g., infiltration, bioretention), or a combination of both. Rev and WQv are inclusive in
that drainage areas having no impervious cover and proposed disturbance may be excluded from
recharge calculations as well. The intent of the recharge requirement is to maintain existing
groundwater recharge at development sites to preserve water table elevations in order to maintain
the hydrology of streams and wetlands under dry-weather conditions. Rev is dependent on slope,
soil type, vegetative cover, precipitation, and evapotranspiration; sites with natural ground cover
exhibit higher recharge rates when compared to impervious surfaces resulting from development.
Channel Protection Volume (Cpv)
The channel protection storage volume(Cpv) requirement exists to protect stream channels from
excessive erosion caused by the increase in runoff from new development. The rationale for this
criterion is that runoff from the 1-yr design storm will be stored and released in such a gradual
manner that critical erosive velocities during bankfull and near-bankfull events will rarely be
exceeded in downstream channels. The Cpv requirement does not apply to direct discharges to
tidal waters.
Of these criteria, the water quality, recharge and channel protection volumes are determined by
soils, amount of imperviousness, proposed design and/or layout, and implementation of
nonstructural practices. This simplifies calculations, reduces error and/or abuse, and provides
direct incentives to reduce impervious areas. Another important feature of these three volumetric
criteria is the relation to natural hydrologic processes. When considered together, these three
criteria capture and treat the runoff from at least 95% of the average annual rainfall and mimic
natural recharge and channel forming processes.
There are two primary approaches for managing stormwater runoff and addressing the unified
stormwater sizing criteria requirements on a development site:
<
the use of better site design practices to reduce the amount of stormwater runoff and
pollutants generated and/or provide for natural treatment and control of runoff, and
<
the use of structural stormwater controls to provide treatment and control of stormwater
runoff (Atlanta Regional Commission, 2001).
Structural stormwater controls should be considered after all reasonable attempts have been made
to minimize stormwater runoff and maximize its control and treatment through better site design
methods. Once the need for structural controls has been established, all relevant stormwater
sizing criteria should be applied in selecting one or more appropriate controls to meet the
stormwater runoff storage and treatment requirements. Most development sites generally require
a combination of structural and/or nonstructural BMPs to meet all stormwater sizing criteria
(WEF and ASCE, 1998).
3-5
3.2.2 BMP Performance Objectives
A fundamental objective of stormwater management programs should be to attempt to reduce the
change from the pre-development hydrology of the site. The use of structural controls for
treating stormwater and to improve the quality of receiving waters has a set of objectives derived
from a number of sources that include: (i) federal, state and local regulatory requirements; (ii)
state or local community goals to mitigate the impacts associated with urban runoff; and, (iii)
special local area needs such as trout or salmon fisheries protection, water supply and watershed
protection, flood control to protect human life and property, groundwater protection, and other
issues of local importance. In general, five different levels of stormwater BMP performance
goals have been identified (Clar et al., 2003).
<
Level 1 - Flood control and peak discharge control
<
Level 2 - Level 1 + 80% TSS removal
<
Level 3 - Flood, peak discharge, and water quality control
<
Level 4 - Unified sizing (multi-parameter) criteria
<
Level 5 - Ecologically sensitive stormwater management
These goals have been discussed in detail in an earlier U.S. EPA publication on structural BMP
design considerations for improving water quality(Clar et al., 2003).
3.3 DESIGN OF STRUCTURAL BMPS TO IMPROVE WATER QUALITY
Stormwater BMPs can be designed for a wide range of goals and objectives, e.g., from a single
parameter approach such as flood control typical in older developed watersheds, or pollutant
removal typical of undeveloped watersheds , to multi-parameter ecological sustainability of
receiving systems. These management goals will determine the requirement for the proper
design and mix of ecological and engineering factors that must be considered. These typically
include hydrology and inflow hydraulics, soil characteristics/infiltration rates, site-specific water
quality concerns, location and site constraints, and the associated costs as well as the condition of
the receiving waters (Clar et al., 2003).
This white paper emphasizes that a clear understanding of the fundamental mechanisms at play
within a BMP to reduce the effluent load is the key to properly designing BMPs. The
preliminary step is the examination of the following criteria: (i) how does the BMP address the
twin issues of stormwater quality and quantity control; (ii) what are the design considerations
involved in achieving optimum water quality control in the BMP; (iii) how effective is the BMP
in terms of meeting performance objectives; (iv) what are the cost factors associated with the
design and construction of the BMP and how to be cost-effective; and more importantly, (v) how
best to place it in the watershed. This white paper focuses on the design considerations of the
following more commonly used structural BMPs :
3.3.1 Dry Extended-detention Ponds
3.3.2 Wet Ponds
3-6
3.3.3 Stormwater Wetland
3.3.4 Grassed Swales
3.3.5 Vegetated Filter Strips
3.3.6 Infiltration Trenches
3.3.7 Porous Pavements
3.3.8 Sand and Organic Filters
The following sections present a detailed discussion on the general design considerations and
BMP design guidelines for these BMPs adopted by various state agencies for stormwater
management. Several well known BMP manuals, (e.g.,Atlanta Regional Commission, 2001;
Schueler, 1987; Florida DER, 1988; NVPDC and ESI, 1992; Schueler, 1992a; WEF and ASCE,
1998; VA DCR, 1999; MDE, 2000; Metro Council, 2001; NYSDEC, 2001; U. S. EPA, 2001;
WA DOE, 2001; Caltrans, 2002; GeoSyntec and ASCE, 2002; CASQA, 2003) have been
consulted in presenting the multitude of approaches practiced by the different agencies in
designing structural BMPs for stormwater management for new and existing urban development.
It must be noted that the ensuing discussion is merely a compilation of the existing design
practices and should not be treated as the ultimate design guidance document. The actual design
process must take into account several locally specific criteria that primarily include the longterm rainfall/hydrological and soil considerations, water quality objectives, economics, and
stormwater regulatory considerations. The typical operation and maintenance requirements for
these BMPs are discussed in chapter 5.
3.3.1 Dry Extended-detention (ED) Ponds
An ED pond is an impoundment that temporarily stores stormwater runoff from a water quality
design storm for a specified minimum period of time (usually 24 to 48 h) and discharges it
through a hydraulic outlet structure to a downstream conveyance system; it is usually dry during
non-rainfall periods and does not have any permanent standing water (CASQA, 2003; VA DCR,
1999). An ED pond can be designed to provide for one or all of the following: water quality
enhancement; downstream flood control; and, channel erosion control. Conventional ED ponds
temporarily detain a portion of stormwater runoff for up to 24 h after a storm using a fixed orifice
resulting in the settling out of urban pollutants; enhanced ED ponds are designed to prevent
clogging and suspension encountered in conventional ED ponds due to frequent high inflow
velocities, and thus have higher efficiencies. ED ponds provide greater flexibility in achieving
target detention times. Along with a detention area, they typically include a sediment forebay
near the inlet, a micropool and/or plunge pool at the outlet, and utilize an adjustable
reverse-sloped pipe as the ED pond control device to prevent resuspension of particles deposited
in earlier storms (WY DEQ, 1999). The detention of runoff for 24 h or more (up to 72 h) helps
to remove up to 90% of particulate pollutants, while the removal of soluble forms of nitrogen and
phosphorus in urban runoff can be enhanced if the normally inundated area of the pond is
managed as a shallow marsh or a permanent pool (Schueler, 1987).
3-7
3.3.1.1 Stormwater Control
ED ponds can be designed for flood control by providing additional storage above the ED
volume, and by reducing the peak rate of runoff from the drainage area (CASQA, 2003). These
BMPs are effective in controlling post-development peak discharge rates to the desired
pre-development levels for the design storm(s) specified, and the optimum level is achieved by
controlling multiple design storms. The design storms are chosen based on ordinance, or
specified watershed conditions (VA DCR, 1999). ED ponds are also capable of managing
smaller floods that contribute to channel erosion problems and occur more frequently than the
annual or 2-yr flood (Schueler, 1987).
3.3.1.2 Pollutant Removal Capability
Pollutant removal is primarily accomplished by gravitational settling that is dependent on the
detention time and the fraction of the annual runoff volume that is effectively detained in the
pond (Schueler, 1987) together with sorption of pollutants to particulates and the associated
settling velocity distribution of these particles. Conventional ED ponds provide moderate but
variable removal of particulate pollutants such as sediment, phosphorus and organic carbon, and
some removal of soluble pollutants (CASQA, 2003). Increasing detention times may result in
greater removal of soluble pollutants. Urban pollutants commonly of concern, including nitrate
and orthophosphates, remain in solution, and may be removed by managing the lower stage of
the ED pond as a shallow wetland that utilize natural biological removal processes (Schueler,
1987; VA DCR, 1999).
Positive factors influencing pollutant removal (WY DEQ, 1999) include:
<
six to twelve hours of minimum detention,
<
smaller treatment volumes (e.g., 0.5 watershed in.) provide the best removal rates,
<
wetlands in lower stage of design prevent resuspension and augment removal of
sediments,
<
use of a micropool to protect the ED pond orifice, and
<
soils that soak up runoff and evapotranspire the same.
Negative factors influencing pollutant removal (WY DEQ, 1999) include:
<
re-suspension of previously deposited pollutants from the pilot channel of pond floor,
<
large treatment volumes (acceptable ED times cannot be achieved over the broad range of
expected storms), and
<
difficulty in predicting ED hydraulics.
3.3.1.3 Design Considerations
General Considerations
The success of an ED pond is dependent on the designer's ability to identify any site and
3-8
downstream conditions that may affect the design and function of the pond. The facility should
be compatible with both upstream and downstream stormwater systems to promote a watershed
approach in providing stormwater management (VA DCR, 1999). The size can be based on the
volume for which BMP credit is desired and the volume being dictated by stormwater
management requirements, as these facilities are usually designed for both water quality and
stormwater management needs. The shape of these facilities is often dictated by site constraints
and topography (WEF and ASCE, 1998).
The dimensions of the pond need to be sized appropriately in order to enhance the effectiveness
of these BMPs. An effective configuration of the pond should result in a long flow path,
promote the establishment of low velocities, and avoid having stagnant areas of the pond. In
order to promote settling of pollutants and aesthetic appeal, the design should consider the
length-to-width ratio, cross sectional areas, pond slopes and configuration, and aesthetics.
Sizing Detention Ponds
There are several ways to size an ED pond. The more common methods use either maximized
volume or hydrograph routing (WEF and ASCE, 1998).
The maximized volume method is the simplest and most direct way for smaller catchments
serving up to approximately 1 km2 (0.6 mi2). The methodology to estimate the maximized water
quality capture volume is described below.
The stormwater quality capture volume may be found by using continuous hydrologic simulation
and local long-term hourly (or lesser increment) precipitation records, or by obtaining a
first-order estimate of the needed capture volume using simplified procedures that target the most
typically occurring population of runoff events. In the U.S., this data is available for ponds that
empty their entire volume in 24 and 48 h. Mean values for other emptying times can be
determined by interpolating the results for the 24 and 48 h time. After an extensive analysis of
the mean annual runoff-producing rainfall depths for the different meteorological regions of the
U.S., simple regression equations were established to relate the mean precipitation depth to
"maximized" water quality runoff capture volumes (WEF and ASCE, 1998). The analytical
procedure was based on a simple transformation of each storm's volume of precipitation to a
runoff volume using a coefficient of runoff. A third order regression equation, Equation 3-1
(WEF and ASCE, 1998), was derived using data from more than 60 urban watersheds (U.S.
EPA, 1983). The equation has broad applicability for smaller storm events in the U.S. as it was
derived from a nationwide monitoring over a 2-yr period.
c ~ 08 58i 3 -
(3-1)
078i J + 0 .774i + 0.04
where
C = runoff coefficient; and
i = watershed imperviousness ratio, namely, percent total imperviousness divided by 100.
3-9
Equation 3-2 relates mean precipitation depth to the "maximized" detention volume and is given
by
Po
~
(3-2)
(a x C) x P6
where
P0 = maximized detention volume determined using either the event capture volume or the
volume capture ratio as its basis, watershed in. (mm);
a = regression constant from least-squares analysis;
C = watershed runoff coefficient; and
P6 = mean storm precipitation volume, watershed in. (mm).
Values of coefficient a have been determined based on an analysis of long-term data from seven
precipitation gauging sites located in different meteorological regions of the U.S. for different
drain times of 12, 24 and 48 h, respectively and shown in Table 3-2 (Guo and Urbonas, 1995).
The correlation of determination coefficient, r2, ranges from 0.80 to 0.97, implying a strong level
of reliability. It is suggested that the event-capture-ratio-based coefficients in Table 3-2 be used
with equation 3.2 instead of the volume capture ratio coefficients (WEF and ASCE, 1998).
While the choice of the emptying or drain time rests with the designer or with the local
authorities, it must be noted that suspended solids are better removed under longer emptying
times. However, the disadvantage of longer drain times is that they tend to produce less
attractive facilities, ones that have little or no vegetation on the bottom, or "boggy" bottoms with
marshy vegetation that pose maintenance concerns.
Table 3-2. Values of Coefficient “a” for Finding the Maximized Detention Storage Volume*
12 h
Drain time of capture volume
24 h
48 h
Event capture ratio
a
r2
1.109
0.97
1.299
0.91
1.545
0.85
Volume capture ratio
a
r2
1.312
0.80
1.582
0.93
1.963
0.85
* Approximately 85 th percentile runoff event (range 82 to 88%)
(Guo and Urbonas, 1995)
The hydrograph routing method is used for detention ponds that serve areas larger than 1 km2 by
converting the maximized storm depth to a design hyetograph, to simulate a runoff hydrograph.
Although the method by which this is done is dictated by the typical design storm temporal
distribution in use within the region where the facility is located, it is suggested that the
maximized depth be redistributed into a 2 h design storm hyetograph. The goal of reservoir
routing is to balance inflow rates against outflow rates to find the needed volume, which can be
3 - 10
accomplished with numerical methods or by using of one of the many available computer
programs written for this purpose (WEF and ASCE, 1998). The needed storage volume is a time
integral of the difference between inflow and outflow hydrographs from the beginning of storm
runoff to the point in time where the outflow rate exceeds the inflow rate (Equation 3-3):
(3-3)
where
Vmax = storage volume;
t = time from beginning of runoff to a point of maximum storage;
Qin = Qout on hydrograph recession limb;
Qin = inflow rate; and
Qout = outflow rate.
Local governments have developed a number of sizing rules for extended-detention, each
specifying both a volume to be detained and a duration over which this volume is released.
(CASQA, 2003) states that capture volume is determined by local requirements or sized to treat
85% of the annual runoff volume. It must be mentioned that a thorough engineering approach to
sizing a pond may be the best option. This can be accomplished by using long-term dry- and
wet-weather flows, pond inflow based on watershed hydrology, and particle settling velocity
calculations together with necessary calculations for pond soil infiltration and evaporation. A
pre-monitoring program to study the settling velocity distribution and the analysis of pollutant
associations with particulates and soluble fraction is recommended.
Siting
Dry ED ponds are among the most widely applicable stormwater management practices and are
especially useful in retrofit situations where their low hydraulic head requirements allow them to
be sited within the constraints of the existing drainage system (CASQA, 2003). The basic
guidelines for siting dry ED ponds are as follows:
<
Dry ED ponds may be used for a wide range of drainage areas. However, the upper range
for contributing drainage area applicable for these ponds without having to take baseflow
into consideration is about 50 to 75 acres (NVPDC and ESI, 1992).
<
Dry ED ponds should be used on sites with a minimum area of 5 acres. With this size
catchment area, the orifice can be on the order of 0.5 in. The challenge in smaller sites is
to provide channel or water quality control because the orifice diameter at the outlet
needed to control relatively small storms becomes very small and is prone to clogging.
<
The base of the extended-detention facility should not intersect the water table, as a
permanently wet bottom may become a mosquito breeding ground (CASQA, 2003).
3 - 11
<
Adequate access from a public or private right-of-way to the pond should be reserved.
The access should be at least 10 ft wide, on a slope of 5:1 or less, and stabilized to
withstand the passage of heavy equipment.
<
All ED ponds should be a minimum of 20 ft from any structure or property line, and 100
ft from any septic tank/drain field. ED ponds should also be a minimum of 50 ft from any
steep slope (greater than 15%). Otherwise, a geotechnical report will be required to
address the potential impact of any pond that must be constructed on or near such a slope
(VA DCR, 1999).
<
ED ponds can be used with almost all soils and geology, with minor design adjustments
for regions of rapidly percolating soils such as sand. In these areas, these BMPs may
need an impermeable layer to prevent groundwater contamination. Highly permeable
soils are not suitable for ED ponds, and for an enhanced ED pond, the soils must support
the shallow marsh at the time of stabilization and planting.
Quantity Detained
The amount of runoff detained heavily influences the pollutant removal performance. At a
minimum, ED ponds should be sized to accommodate the runoff produced by the mean storm,
and preferably should be capable of storing the runoff volume of a 1.0 in. storm. Higher levels of
control can be achieved when the runoff volume from the 1- or 2- yr storm is detained (Schueler,
1987). However, in many cases, the stricter storage requirements recommended above for
streambank erosion control ( 1.0 to 1.5 in. Rv) will govern how much extra detention storage is
needed.
Duration
Detention times of at least 24 to 36 h are probably necessary to achieve maximum removal of
most pollutants. Although most of the settling occurs within the first 12 h in settling column
experiments, it is advisable to provide further detention since several hours may be needed before
ideal settling conditions develop in a pond. Slightly longer detention times of up to 40 h may be
needed in larger watersheds for downstream channel erosion control. The control device must be
sized so as to provide an adequate detention time for the entire spectrum of storms. The pond
designer should perform several storage routing calculations e.g., TR-20 method or equivalent, to
determine the approximate detention time for the smaller, more frequent runoff events. As a
general rule, it is recommended that the average detention time for small runoff events (0.1 to 0.2
in.) should be no less than 6 h. As a final check, the runoff velocity of the downstream channel
at the extended-detention release rate should be computed to make sure that it is not erosive
(Schueler, 1987).
Pond Configuration
Minimizing the velocity of the flow through the pond greatly improves the pollutant removal
3 - 12
efficiency of the pond, which can be effected by increasing the pond depth as well as the cross
sectional area (NVPDC and ESI, 1992). The basin should gradually expand from the inlet and
contract toward the outlet to reduce short circuiting and slow influent velocities by increasing the
cross sectional flow area. The goal is to provide conditions where the velocity of flow through
the facility for a typical storm event is less than the settling velocities of the pollutants of concern
(NVPDC, 1979).
The length-to-width ratio of a pond is one design aspect that can significantly affect pollutant
removal. The distance between inlet and outlet points needs to be maximized in order to
promote pollutant settling. A high aspect ratio may improve the performance of detention ponds;
consequently, the outlets should be placed to maximize the flowpath through the facility. While
the flow path length is defined as the distance from the inlet to the outlet as measured at the
surface, the average width is calculated as the surface area of the pond divided by the length
(NVPDC and ESI, 1992) (Metro Council, 2001). A length-to-width ratio of two or greater,
preferably up to a ratio of four is required for additional detention time for settling and biological
treatment (WEF and ASCE, 1998).
Pond depths optimally range from 2 to 5 ft and may include a sediment forebay to provide the
opportunity for larger particles to settle out (NVPDC and ESI, 1992).
A micropool is not recommended in the design because of vector concerns. For online facilities,
the principal and emergency spillways must be sized to provide 10 ft of freeboard during the
25-yr event and to safely pass the flow from 100-yr storm (CASQA, 2003).
Pond Side Slopes
Pond side slopes need to be stable under saturated soil conditions. They also need to be
sufficiently gentle to limit rill erosion, facilitate maintenance, and address the safety issue of
individuals falling in when the basin is full of water. In order to promote facility effectiveness, it
is highly desirable to avoid resuspension of materials collected on the pond floor; the potential
for resuspension is generally minimized by reducing inflow velocities and maintaining vegetative
cover. Side slopes should be no steeper than 4:1 (H:V), and no flatter than 20:1 (H:V) (Schueler,
1987; WEF and ASCE, 1998). Slopes steeper than this needs to be stabilized with an appropriate
slope stabilization practice (CASQA, 2003).
Pond Lining
Ponds must be designed to prevent possible contamination of groundwater below the facility.
Pond Inlet
An ideal inflow structure should convey stormwater to the pond while preventing erosion of the
pond bottom and banks, reducing resuspension of previously deposited sediment, and facilitating
deposition of the heaviest sediment near the inlet. Such energy dissipation measures also reduce
3 - 13
the tendency for short-circuiting. Inflow structures can be drop manholes, rundown chutes with
an energy dissipator near the bottom, a baffle chute, a pipe with an impact basin, or one of the
many other types of diffusing devices, depending on pond geometry (NVPDC and ESI, 1992;
WEF and ASCE, 1998).
Outflow Structure
The outlet should be capable of slowly releasing the design capture volume over the design
emptying time. ED ponds are designed to encourage sediment deposition and as stormwater has
substantial quantities of settleable and floatable solids, outlets are prone to be clogged,
invalidating the hydraulic function of even the best design. Each outlet therefore needs to be
designed with clogging, vandalism, maintenance, aesthetics, and safety in mind (WEF and
ASCE, 1998). The facility's drawdown time should be regulated by a gate valve or orifice plate.
In general, the outflow structure should have a trash rack or other acceptable means of preventing
clogging at the entrance to the outflow pipes. The structure should be sized in such a way to
allow for complete drawdown of the WQv in 72 h, with no more than 50% of the water quality
volume draining from the facility within the first 24 h. The outflow structure should be fitted
with a valve in order to regulate the rate of discharge from the basin as well as to halt the
discharge in case of an accidental spill in the watershed (CASQA, 2003). The discharge from a
control orifice is given by Equation 3-4:
(3-4)
where
Q = discharge (ft3/s);
C = orifice coefficient;
A = area of the orifice (ft2);
g = gravitational constant (32.2 ft/s2);
H = water surface elevation (ft); and
H0 = orifice elevation (ft).
Recommended values for C are 0.66 for thin materials and 0.80 when the material is thicker than
the orifice diameter. This equation can be used with the pond stage/volume relationship to
calculate drain time in spreadsheet form.
Storage Volume
The storage volume in the ED pond should be equal to the maximized volume discussed earlier.
An additional 20% could be added to this volume to provide for sediment accumulation, and
could be used to promote sedimentation of smaller particles (less than 60 µm in size), which
account for approximately 80% of the suspended sediment mass found in stormwater (WEF and
ASCE, 1998)
3 - 14
Flood Control Storage
Whenever feasible, the ED pond should be incorporated within a larger flood control facility. By
doing so, both water quality and flood control functions can be combined in a single detention
basin.
Two-Stage Design
This pond configuration is meant to address both water quality and quantity. A two-stage basin
is recommended when extended-detention is applied to dry ponds. The top stage of the pond
should have the capacity to regulate peak flow rates of large infrequent storms (10-, 25- or 100­
yr), and will generally remain dry between storms. The volume in this stage is called the "flood
storage volume." The second stage of the pond is designed to detain smaller storms for a
sufficient period of time to remove pollutants from the runoff. The volume in this stage is called
the "water quality volume." For ED ponds, WQv is typically the runoff from the 0.3-yr storm
event, since a large fraction of the annual pollutant load is delivered by small, frequent storm
events (Metro Council, 2001).
The upper stage of the pond is sized and graded (2% minimum) to remain dry except during
large infrequent storms, while the bottom stage is expected to be regularly inundated. The lower
portion has a micropool that fills frequently and reduces the periods of standing water and
sediment deposition in the remainder of the basin. A marsh-like environment in the lower
section allows for some biological uptake of soluble materials and provides quiescent conditions,
which promotes sedimentation of particulates (NVPDC and ESI, 1992). However, these
recommendations do not necessarily apply to large, regional extended-detention ponds (WEF and
ASCE, 1998). Extra storage, over and above stormwater and extended-detention requirements,
should be provided within the bottom stage, or at the inlet to account for 20 yr of sediment
deposition. The main advantage of this configuration is that frequently inundated areas are
localized in one section of the pond, thus allowing the upper portion of the BMP facility to be
used for certain low intensity recreational uses during dry weather.
Enhancement Options - Wetland Creation
Establishing wetland vegetation in a shallow marsh component or on an aquatic bench in the
lower stage of the detention basin will enhance removal of soluble nutrients, increase sediment
trapping, prevent sediment resuspension, and provide wildlife and waterfowl habitat (Metro
Council, 2001). The use of a shallow marsh limits the maximum range of vertical storage in the
ED pond to 3 ft above the marsh's water surface elevation. However, the surface area
requirements for the shallow marsh will likely force the basin's geometry to broaden at the lower
stages, which will compensate for the vertical storage (VA DCR, 1999). In general,
extended-detention water surface elevations greater than 3 ft, and the frequency at which those
elevations can be expected, are not conducive to the growth of dense or diverse stands of
emergent wetland plants. Water depths of 6 to 12 in. would be required for optimal wetland
growth, and native species should be planted in the wetland (Schueler, 1987). The general
3 - 15
guidelines in designing a shallow marsh are discussed in detail in section 3.3.3 on stormwater
wetlands.
Sediment Forebay
The settling area for incoming sediments can be increased through the addition of a sediment
forebay. The use of a sediment forebay, however, is only recommended for wet ponds larger
than 4,000 ft3. The forebay is an excavated settling basin or a section separated by a low weir at
the head of the primary impoundment. Forebays serve to trap sediments before the sediment
enters the primary pool, effectively enhancing removal rates and minimizing long-term operation
and maintenance problems. Also, it is easier and more cost effective to remove sediments from
the forebay periodically as compared to removal from a wet pond pool. Hard bottom forebays
make sediment removal easier, and forebays should be accessible to heavy machinery, if
necessary. About 10 to 25% of the surface area of the wet pond should be devoted to the
forebay.
The forebay can be distinguished from the remainder of the pond by one of several means: a
lateral sill with rooted wetland vegetation; two ponds in series; differential pool depth;
rock-filled gabions or retaining walls; or, a horizontal rock wall filter placed laterally across the
pond. Energy dissipation techniques should be used at the inlet to the sediment forebay to avoid
erosion, promote settling, and minimize short-circuiting of flows. The length-to-width ratio of
the forebay should be at least 2:1 to minimize short-circuiting (Metro Council, 2001).
Low Flow Channels
Low flow channels route the last remaining runoff, dry weather flow and groundwater to the
permanent pool and outlet. These channels should be installed in the upper stage of the basin to
ensure that the basin dries out completely. Low flow channels also serve to prevent erosion of
the upper stage of the pond outside as runoff first enters the pond.
The presence of a baseflow makes the design of an extended-detention control structure difficult.
An orifice designed for wet-weather baseflow, compromises the dry-weather control due to very
high release rates. On the contrary, if an orifice is undersized to meet the dry-weather control,
the ED pond may remain full of water during the wet-weather season and eliminate the
extended-detention volume by creating an undersized permanent pool (VA DCR, 1999). When
seasonal baseflow is present, an adjustable orifice should be provided in the control structure to
maintain the marsh volume.
Design considerations should take into account the presence of a baseflow and the associated
potential for erosion within the basin and, ideally, spread them out so they sheet flow across the
bottom of the basin. A few local ordinances require the use of low-flow channels to carry
baseflows. Generally, an impervious low-flow channel is not recommended in a stormwater
management water quality basin, as its use is contrary to the basin's water quality function.
However, an impervious ditch may be used to carry baseflow if it is designed to overflow during
3 - 16
storm events and spread the runoff across the basin floor. The use of gabion baskets or riprap,
instead of concrete, may provide the advantage of slowing the flow, encouraging spillover onto
the basin floor.
Overflow
Similar to a constructed stormwater wetland, an ED overflow system should be designed to
provide adequate overflow or bypass for a full range of design storms.
Liner to Prevent Infiltration
ED ponds should have negligible infiltration rates through the bottom of the pond. If infiltration
is anticipated, and the area is not suspected to be underlain by karst, then an infiltration facility,
rather than a detention water quality BMP should be used or a liner should be installed in the
basin to prevent infiltration. The following recommendations apply when using a liner:
<
A clay liner should have a minimum thickness of 12 in. and should comply with the
specifications provided in Table 3-3.
<
A layer of compacted soil (minimum 6 to 12 in. thick) should be placed over the liner
before seeding with an appropriate seeding mixture.
<
Other liner types may be used if supporting documentation is provided verifying the liner
material's performance.
Table 3-3. Clay Liner Specifications
Property
Test Method
(or equal)
Unit
Specification
Permeability
ASTM D-2434
cm/sec
1 x 10 -6
Plasticity Index of
Clay
ASTM D-423 & D­
424
%
Not less than 15
Liquid Limit of Clay
ASTM D-2216
%
Not less than 30
Clay Particle Sieving
ASTM D-422
%
Not less than 30
Clay Compaction
ASTM D-2216
%
95% of Standard
Proctor Density
(City of Austin, 1988)
Pond Buffer
A buffer strip away from the pond to the nearest lot should be reserved and landscaped using
low-maintenance grasses, shrubs and trees (e.g., the minimum width should be 25 ft (Schueler,
1987). A landscaping plan for the pond and buffer should outline measures to improve the
3 - 17
appearance for adjacent residents, meet specific design functions, and provide local wildlife
habitat (Schueler, 1987).
Dam Embankment
The dam embankment should be designed not to fail during storms larger than the water quality
design storm. An emergency spillway could be provided, the design of which is governed by
local regulations, and the embankment should have at least one foot of freeboard above the
emergency spillway. The other approach is to design the embankment to withstand overtopping
commensurate with embankment size, the volume of water that can be stored in it, and the
potential of downstream damages or loss of life if the embankment fails. Embankments for small
onsite basins should be protected from at least the 100-yr flood, while the larger facilities should
be evaluated for the probable maximum flood. Embankment slopes should be no steeper than
3:1, preferably 4:1 or flatter. They also need to be planted with turf-forming grasses.
Embankment soils should be compacted to 95% of their maximum density at optimum moisture,
graded to allow access for heavy equipment, and mowed twice a year to prevent woody growth.
At least 10 to 15% extra fill should be allowed on the embankment to account for possible
subsidence.
Vegetation
Vegetation provides erosion control and enhances sediment entrapment in a pond. The pond can
be planted with native grasses or with irrigated turf, depending on the local setting, pond design,
and its intended other uses such as recreation. The maintenance of a healthy grass cover on the
pond bottom is difficult due to sediment deposition, along with frequent and prolonged periods
of inundation. Options for an alternative bottom liner include a marshy wetland bottom, bog,
layer of gravel, riparian shrub, bare soil, low-weed species, or other type that can survive the
conditions existing in the bottom of the pond.
Splitter Box
Splitter structures isolate WQv, when ponds are designed as offline facilities. The splitter box, or
other flow diverting approach, should be designed to convey the large storm event (e.g., 25- yr
storm) while providing adequate (e.g., at least 10 ft) freeboard along pond side slopes.
Erosion Protection at the Outfall
For online facilities, special consideration should be given to the facility's outfall location.
Flared pipe end sections that discharge at or near the stream invert are preferred. The channel
below the pond outfall should be modified to conform to natural dimensions, and lined with large
stone riprap placed over filter cloth. Energy dissipation may be required to reduce flow
velocities from the primary spillway to non-erosive velocities. An example schematic of a dry
ED pond (MDE, 2000) is shown in Figure 3-1.
3 - 18
~EMERGENCY
SPILLWAY
INFLOW
PLAN VIEW
EMBANKMENT
v
100 YEAR lEVEL
sz
_
10 YEAR lEVEL
~O
CPo or 2 YEAR lEVEL
wa
·EO ElEVATIO
IIJ)""
. •'
'!!Im '-liit~dii~W:u~:::;
FOREBAY
M1CROPOOl
"""'EC
ANTI-SEEP COl1.AR
FilTER OIAPHRAGM
OI--.!
"
PROFILE
Figure 3-1. Schematic of a Dry Extended-Detention Pond (MDE, 2000)
3 - 19
3.3.2 Wet Ponds
A wet pond is a constructed stormwater retention basin with emergent wetland vegetation around
the perimeter. It is designed to have a permanent pool of water throughout the year or at least
during the wet season. Runoff from each rain event is detained and treated in the pool primarily
through gravitational settling and biological uptake mechanisms until it is displaced by runoff
from the next storm (Atlanta Regional Commission, 2001). The permanent pool provides a
vessel for the settling of solids between storms and the removal of nutrients and dissolved
pollutants. The wetland vegetation, also called the littoral zone, provides aquatic habitat,
enhances pollutant removal, and reduces the formation of algal mats, and can be created by
excavating an already existing natural depression or through the construction of embankments.
Conventional wet ponds have a permanent water pool to treat incoming stormwater runoff. In
enhanced wet pond designs, a forebay is installed to trap incoming sediments where they can be
easily removed; a fringe wetland is also established around the perimeter of the pond (WY DEQ,
1999). .
3.3.2.1 Stormwater Control
Stormwater ponds are designed to control both stormwater quality and quantity and can be used
to address all the unified stormwater sizing criteria for a given drainage area. Wet ponds can be
effective in controlling post-development peak discharge rates to pre-development levels for
desired design storms. Groundwater recharge in wet ponds is limited to the storage lost to
infiltration through the pond bottom. Although the quantity of recharge is greater than that
achieved in dry or extended detention ponds, it is negligible in comparison to infiltration and
other volume control BMPs. The post-development increase in the total runoff volume from a
site is not effectively modified by wet ponds. While some temporary control of runoff volume
happens when extra dead-storage is created by evaporation or infiltration, it generally occurs
during minor storms in the summer months and after prolonged droughts.
3.3.2.2 Pollutant Removal Capability
Pollutant removal in wet ponds is highly variable from storm to storm but generally high over the
long-term, for well designed and maintained ponds. The degree of pollutant removal achieved by
a pond is a function of the size and design of the permanent pool and the characteristics of
individual urban pollutants. Suspended sediments in stormwater runoff settle out from the water
column to the pond sediments. The permanent pool additionally acts as a barrier to resuspension
of deposited materials and improves removal performance over that achieved by dry ponds. The
greatest initial settling often occurs near the pond inlet under quiescent conditions; settling can be
modeled assuming Stokes Law Type I sedimentation. However, pollutant removal rates may
decline during larger storms in smaller ponds due to short-circuiting and the volume of incoming
runoff being greater than the volume of the permanent pool.
A unique feature of wet ponds is the presence of aquatic plants and algae that can remove
3 - 20
significant amounts of soluble nutrients from the water column; retention ponds can be superior
to ED ponds for the control of dissolved nutrients in stormwater (Schueler, 1987). Since soluble
nutrients have minimal settling velocities, biological uptake represents an important removal
pathway. Retention ponds are most appropriate where nutrient loadings are of concern,
especially in the following situations:
<
Watersheds tributary to reservoirs and lakes - retention ponds in the watershed can help
achieve eutrophication management goals in downstream reservoirs and lakes.
<
Watersheds tributary to tidal embayments and estuaries - nutrient loadings into estuarine
systems is a growing concern in coastal areas, including upland areas that drain into tidal
waters; retention ponds can help reduce the nutrient loads.
The degree of pollutant removal is a function of pool size in relation to contributing watershed
area, and is achieved by gravitational settling, algal settling, wetland plant uptake, and bacterial
decomposition. Unlike ED ponds, wet ponds avoid resuspension, and nutrient cycling in these
ponds is generally thought to operate much as in natural lakes; consequently, the pollutant
removal capabilities can be successfully predicted by applying a controlled lake eutrophication
model. The principal factors governing nutrient cycling are the loading and the decay rates for
phosphorus, hydraulic residence time, and mean depth (NVPDC and ESI, 1992).
The observed pollutant removal of a wet pond is highly dependent on two factors; i.e., the
volume of the permanent pool relative to the amount of runoff from the typical event in the area
and the quality of the baseflow that sustains the permanent pool. If the permanent pool is much
larger than the volume of runoff from an average event then the primary process is the
displacement of the permanent pool by the wet-weather flow (Caltrans, 2002). The discharge
quality of wet ponds during dry- and wet-weather flows is not significantly different, resulting in
a relatively constant discharge quality during storms that is the same as the concentrations
observed in the pond during ambient (dry weather) conditions and so are better characterized by
the average effluent concentration, rather than the "percent reduction" (CASQA, 2003). The dryand wet-weather discharge quality is thus related to the quality of the baseflow that sustains the
permanent pool and the transformations of those pollutants during their residence in the basin.
Positive factors influencing pollutant removal include:
<
pretreatment by sediment forebay,
<
permanent pool, 0.5 to 1.0 in. per impervious acre treated,
<
fringe wetlands,
<
shallow wetlands and/or extended detention may improve removal efficiencies, and
<
high length-to-width ratios.
Negative factors influencing pollutant removal include::
<
small pool size,
<
fecal contribution from large waterfowl populations,
<
short-circuiting and turbulence,
<
sediment phosphorus release,
3 - 21
<
<
extremely deep pool depths (greater than 10 ft), and
snowmelt conditions and/or ice.
3.3.2.3 Design Considerations
A well designed stormwater pond consists of the following:
<
permanent pool of water,
<
overlying zone in which runoff control volumes are stored, and
<
shallow littoral zone (aquatic bench) along the edge of the permanent pool that acts as a
biological filter (WEF and ASCE, 1998).
Siting
Wet ponds are a widely applicable stormwater management practice and can be used over a
broad range of storm frequencies and sizes, drainage areas, and land use types. They can be
constructed on- or off-line (off-line is preferred) and can be sited at feasible locations along
established drainage ways with consistent baseflow. Wet basin application is appropriate in the
following settings:
<
<
<
<
where there is a need to achieve a reasonably high level of dissolved contaminant removal
and/or sediment capture,
in small to medium-sized regional tributary areas with available open space and drainage
areas greater than about 10 ha (25 acre),
where baseflow rates or other channel flow sources are relatively consistent year-round,
and
in residential settings where aesthetic and wildlife habitat benefits can be appreciated and
maintenance activities are likely to be consistently undertaken (CASQA, 2003).
Soil Permeability
Highly permeable soils may not be acceptable for retention ponds because of excessive
drawdown during dry periods. Where permeable soils are encountered, exfiltration rates can be
minimized by scarifying and compacting a 0.3 m (12 in.) layer of the bottom soil of the pond,
incorporating clay to the soil, or providing an artificial liner. Excavating the permanent pool into
the groundwater table can also ensure its permanency, but seasonal fluctuations in the
groundwater table need to be taken into account (Schueler, 1987).
Design Criteria
The design of permanent pools for a wet pond employs two different methods.
<
The solids-settling design method relies on the solids-settling theory and assumes that all
pollutant removal is because of sedimentation.
<
The lake eutrophication model design method provides for a level of eutrophication by
accounting for the principal nutrient removal mechanisms (WEF and ASCE, 1998).
3 - 22
The solids-settling method is most appropriate for situations where the control of total suspended
sediments and pollutants that attach themselves to the solids is the principal objective. The
method relies on rainfall and runoff statistics, pond size, and settling velocity distributions of
suspended solids to calculate total suspended sediment removal. This method assumes an
approximate plug flow system in the retention pond with all pollutant removal resulting from
sedimentation. Testing this model using data from nine retention ponds monitored during U.S.
EPA's NURP showed that it predicted removal rates reasonably well (WEF and ASCE, 1998).
The lake eutrophication model assumes that a retention pond is a small eutrophic lake that can be
represented by empirical models used to evaluate lake eutrophication effects. This method is
used to size a retention pond to achieve a controlled rate of eutrophication and an associated
removal rate for nutrients. Retention ponds that achieve nutrient removal also removes other
pollutants, and typically it is not necessary for the design process to address constituents other
than nutrients.
Like most input/output lake eutrophication models, this model is an empirical approach that
treats the permanent pool as a completely mixed system and assumes that it is not necessary to
consider the temporal variability associated with individual storm events. While the
solids-settling model accounts for the temporal variability of individual storms, the lake
eutrophication model is based on annual flows and loadings.
The model is applied in two parts:
K=
O.56xQ,
Fx(Q, +13.3)
(3-5)
where
K2 = second order decay rate m3/mg x a;
Qs = Z/T the mean overflow rate, m/a;
Z = mean pond depth, m;
T = average hydraulic retention time, yr; and
F = inflow (ortho P/total P) ratio.
R=lOt10
,Jl0+(10+4N)
2N
(3-6)
where
R = total P retention coefficient, (i.e. BMP efficiency);
N = K2 x PT x T; and
PT = inflow total P, µg/L.
These two equations were developed from a database for 60 U. S. Army Corps of Engineers'
3 - 23
reservoirs and were verified for 20 other reservoirs. When this model was applied to 20 other
reservoirs, 10 NURP sites and 14 other retention pond systems and small lakes, the
goodness-of-fit test yielded an R2 = 0.8, indicating a good job of replicating monitored total P
removal (WEF and ASCE, 1998).
Sizing
State and regional stormwater management regulations and guidelines often address design
criteria for the permanent pool storage volume in terms of either the average hydraulic retention
time, or minimum total suspended sediment removal rate. The size of the permanent pool in
relation to the contributing watershed is perhaps the single largest factor influencing pollutant
removal in wet ponds. A number of wet pond sizing rules that variously specify the minimum
volume of the permanent pool have been proposed to optimize pollutant removal. There is no
individual rule that can be recommended as applicable in all cases, and the choice in many cases
rests with local stormwater management policy makers. Sizing should take into account the
objective that the pond should be sized to hold the permanent pool as well as the required water
quality volume. An example (Schueler, 1987) is provided in Table 3-4. From the table, it can be
seen that the choice of an appropriate pond sizing rule necessarily invites a trade-off between the
degree of removal efficiency desired and the cost of achieving it.
Table 3-4. Summary of Wet Pond Sizing Rules
Sizing Rule
Sediment
removed
Phosphorus
removed
Extra storage
Extra cost
(compared to 2 yr dry pond)
60-90%
RULE 1:
0.5 in. runoff per
acre
35-90%
35-200%
20-90%
60%
RULE 2:
0.5 in. runoff per
impervious acre
35-40%
30%
20-25%
RULE 3:
0.1 to 0.8 in.
depending on
land use
55-80%
30-50%
30-70%
20-40%
RULE 4:
2.5 times the
runoff of the
mean storm
75%
55%
75%
40-50%
3 - 24
RULE 5:
4.0 times the
runoff of the
mean storm
(app. 2 week
retention)
(Schueler, 1987)
85-90%
65%
200-250%
80-100%
Pond Shape and Geometry
The wet basin should be configured as a two-stage facility with a sediment forebay and a main
pool. Long, narrow and irregular shapes are also desirable for shallower ponds since they reduce
surface area exposed to the wind and thereby prevent resuspension of previously deposited
materials (Schueler, 1987). The basin should be wedge-shaped, narrowest at the inlet and widest
at the outlet (CASQA, 2003).
The minimum length-to-width ratio for the permanent pool shape is 1.5:1, and should ideally be
greater than 3:1 to avoid short-circuiting. Baffles, pond shaping or islands can be added within
the permanent pool to increase the flow path (Atlanta Regional Commission, 2001).
Mean depth of the permanent pool is calculated by dividing the storage volume by the surface
area. The mean depth should be shallow enough to ensure aerobic conditions and reduce the risk
of thermal stratification but deep enough to ensure that algal blooms are not excessive and reduce
resuspension of settled pollutants during significant storm events.
The minimum depth of the open water area should be greater than the depth of sunlight
penetration to prevent emergent plant growth in this area, namely, on the order of 2 to 2.5 m (6 to
8 ft). A mean depth of approximately 1 to 3 m (3 to 10 ft) should produce a pond with sufficient
surface area to promote algal photosynthesis and should maintain an acceptable environment
within the permanent pool for the recommended average hydraulic retention times, although
separate analyses should be performed for each locale (WEF and ASCE, 1998).
If the pond has more than 0.8 ha (2 acre) of water surface, mean depths of 2m (6.5 ft) will protect
it against wind generated resuspension of sediments. A water depth of approximately 1.8 m (6
ft) over the major portion of the pond will also increase winter survival of fish (Schueler, 1987).
A maximum depth of 3 to 4 m (10 to 13 ft) should reduce the risk of thermal stratification;
however, in the state of Florida, pools up to 9.2 m (30 ft) deep have been successful when
excavated in high groundwater areas, probably because of improved circulation at the bottom of
the pond as a result of the movement of groundwater through it.
The perimeter of all permanent pool areas with depths of 4 ft or greater should be surrounded by
an aquatic bench that extends inward 5 to 10 ft from the perimeter of the permanent pool and
should be no more than 18 in. below normal depth. The area of the bench should not exceed
about 25% of pond surface. The depth in the center of the basin should be 4 to 8 ft deep to
3 - 25
prevent vegetation from encroaching on the pond open water surface.
Side Slopes Along Shoreline and Vegetation
Side slopes along the shoreline of the retention pond should be 4H:1V. CASQA (2003)
recommends 3:1 or flatter to facilitate maintenance and reduce public risk of slipping and falling
into the water. Additionally, a littoral zone should be established around the perimeter of the
permanent pool to promote the growth of emergent vegetation along the shoreline and deer
populations from wading. This bench for emergent wetland vegetation should be at least 3 m (10
ft) wide with a water depth of 0.15 to 0.45 m (0.5 to 1.5 ft). The total area of the aquatic bench
should be 25 to 50% of the permanent pool's water surface area. The use of wetland vegetation
within shallow sections of the permanent pool should adhere to guidelines issued by local
agricultural agencies or commercial nurseries. Emergent plants such as bulrush, three-square and
lizards tail can provide an attractive fringe habitat (Schueler, 1987), providing food and cover for
wildlife and waterfowl.
Extended-detention Zone Above the Permanent Pool
Some state or local regulations require detention of a specified runoff volume as surcharge above
the permanent pool, in order to reduce short-circuiting and enhance settling of total suspended
sediments. Although the addition of an extended-detention zone above the permanent pool may
not likely produce measurable increases in the removal of total suspended sediments, it is still
recommended to have a surcharge extended detention volume, and whenever one is used or
required, it is suggested in these local guidelines that the maximum event-based volume with a
12 h drain time be used..
Minimum and Maximum Tributary Catchment Areas
Stormwater ponds should have a minimum contributing drainage area of 25 acres or more for a
wet pond to maintain a permanent pool. The minimum drainage area should permit sufficient
baseflow to prevent excessive retention times or severe drawdown of the permanent pool during
dry seasons. It is recommended that a water balance calculation be performed using local runoff,
evapotranspiration, exfiltration, and baseflow data to ensure that the baseflow is adequate to keep
the pond full during the dry season. The maximum tributary catchment area should be set to
reduce the exposure of upstream channels to erosive stormwater flows, reduce effects on
perennial streams and wetlands, and reduce public safety hazards associated with dam height
(WEF and ASCE, 1998).
Construction of Retention Ponds in Wetland Areas
Although wet pond BMPs are typically designed to enhance pollutant removal by incorporating
wetland areas along the perimeter, regulatory agencies may restrict their use if a significant
amount of native wetlands will be submerged within the permanent pool. If field inspections
indicate that a significant wetlands area will be affected at a particular site, and if the
3 - 26
construction of a wet pond is inevitably required, the following options may be pursued during
final design subject to approval: investigate moving the embankment and permanent pool
upstream of the major wetland area; and if this is not feasible, a wetland mitigation plan can be
developed as a part of the retention pond design.
Forebay
A sediment forebay with a hardened bottom should be constructed near the inlet to trap coarse
sediment particles in order to reduce the frequency of major clean-out activities within the pool
area. The forebay storage capacity should be approximately 10% of the permanent pool storage
and should be at least 3 ft deep (WEF and ASCE, 1998; CASQA, 2003). Exit velocities from the
forebay should not be erosive. A fixed vertical sediment depth marker should be installed in the
forebay to measure sediment accumulation. Access for mechanized equipment should be
provided to facilitate removal of sediment. The forebay can be separated from the remainder of
the permanent pool by one of several means: a lateral sill with wetland vegetation; two ponds in
series; differential pool depth; rock-filled gabions; a retaining wall; or a horizontal rock filter
placed laterally across the permanent pool.
Inlet and Outlet Structures
The inlet design should dissipate flow energy and diffuse the inflow plume where it enters the
forebay or permanent pool. Examples of inlet designs include drop manholes, energy dissipaters
at the bottom of paved rundowns, a lateral bench with wetland vegetation, and the placement of
large rock deflectors.
An outlet for a retention pond typically consists of a riser with a hood or trash rack to prevent
clogging and an adequate anti-vortex device for basins serving large drainage areas. A few
examples are outlet works with surcharge detention for water quality, negatively sloped pipe
outlet with riser, and multiple orifice outlet. Anti-seep collars should be installed along outlet
conduits passing through or under the dam embankment. If the pond is part of a larger
peak-shaving detention basin, the outlet should be designed for the desired flood control
performance. An emergency spillway must be provided and designed using accepted engineering
practices to protect the basin's embankment. The pond embankment and spillway should be
designed in accordance with federal, state, and local dam safety criteria. For on-line facilities,
the principal and emergency spillways must be sized to provide 1.0 ft. of freeboard during the 25­
yr event and to safely pass the 100- yr flood (CASQA, 2003). The channel that receives the
discharge from the basin's outlet should be protected from erosive discharge velocities. Options
include riprap lining of the channel or providing stilling basins, check dams, rock deflectors, or
other devices to reduce outfall discharge velocities to nonerosive levels.
When the pond is designed as an off-line facility, a splitter structure is used to isolate the water
quality volume. The splitter box, or other flow diverting approach, should be designed to convey
the 25- yr event while providing at least 1.0 ft of freeboard along pond side slopes.
3 - 27
Each pond must have a bottom drain pipe with an adjustable valve that can completely or
partially drain the pond within 24 h. However, this requirement may be waived for coastal areas,
where positive drainage is difficult to achieve due to very low relief (Atlanta Regional
Commission, 2001). The pond drain pipe should be sized one pipe size greater than the
calculated design diameter. The drain valve is typically a handwheel-activated knife or gate
valve. Valve controls shall be located inside of the riser at a point where they (i) will not
normally be inundated, and (ii) can be operated in a safe manner. Figure 3-2 provides an
example schematic of a wet pond (CASQA, 2003)
3 - 28
~EMERGENCY
sPILLWAY
INFLOW
FORESAY
RISER IN
EM6AHKMENT
MAINTENANCE
ACCESS ROAD
MAXIMUM ED LIMIT
MAXIMUM SAFETY STORM LIMIT
J
_--------\"""'''''"""'CE-n-'-ECNC--H-PLAN VIEW
-
I NFLOW
FOflEBAY
",ii-, iijlm1]llII;dITlL
U""'j"I=~lt:I!11
E
PROFILE
Figure 3-2. Schematic of a Wet Pond (CASQA, 2003)
3 - 29
3.3.3 Stormwater Wetlands
Stormwater wetlands can be defined as constructed wetland systems that are explicitly designed to
mitigate the impacts of stormwater quality and quantity that occur during the process of
urbanization. They do so by temporarily storing stormwater runoff in shallow pools that create
growing conditions suitable for emergent and riparian wetland plants and routing runoff through
vegetation to maximize contact. The runoff storage, complex micro topography and emergent
plants in the stormwater wetland together form an ideal matrix for the removal of urban pollutants
(Schueler, 1992a). Stormwater wetlands have been characterized as having one of five basic
designs: (1) shallow marsh system; (2) pond/wetland system; (3) extended detention wetland; (4)
pocket wetlands; and, (5) fringe wetlands.
Conventional stormwater wetlands are shallow pools that create growing conditions suitable for
the growth of marsh plants. These are constructed systems and typically are not located within
delineated natural wetlands. Stormwater wetlands differ from other artificial wetlands created to
comply with mitigation requirements in that they may not replicate all the ecological functions of
natural wetlands. However, as with natural wetlands, stormwater wetlands require a continuous
baseflow or a higher table to support aquatic vegetation (Atlanta Regional Commission, 2001).
Functional differences depend on the design of the wetland, interactions with groundwater and
surface water, and local storm climate (WY DEQ, 1999). Enhanced stormwater wetlands
designed for more effective pollutant removal and species diversity also include design elements
such as a forebay, complex microtopography, and pondscaping with multiple species of wetland
trees, shrubs and plants.
3.3.3.1 Stormwater Control
Constructed stormwater wetlands should generally not be used for flood control or stream channel
erosion control due to the anticipated water level fluctuations associated with quantity controls
(VA DCR, 1999). The clearing of vegetation and the addition of impervious surfaces may cause
large and sudden surges of runoff during rain events, and may cause less than normal baseflows
during dry periods. Large, sudden fluctuations in water levels can stress emergent wetland and
upland edge vegetation, most of which cannot survive drought or saturation extremes, leaving
wetland banks exposed to potential erosion. The large surface area requirement for constructed
stormwater wetlands will help to minimize the "extreme" water level fluctuations during all but
the larger storm events and the wetland design should allow for gradual increases and increases in
wetland design.
3.3.3.2 Pollutant Removal Capability
Wetlands remove pollutants through gravitational settling, wetland plant uptake, adsorption,
physical filtration and microbial decomposition. Primary removal of stormwater pollutants occurs
during the relatively long quiescent period between storms (WY DEQ, 1999) and the degree of
pollutant removal is a function of aquatic treatment volume, surface area-to-volume ratio, and the
3 - 30
ratio of wetland surface area to watershed area. Pollutant removal is also expected to increase
with longer storm water flow paths through the wetland and longer residence time within the
wetland. Conventional stormwater wetlands have a high pollutant removal capability that is
generally comparable to that of wet ponds. While sediment removal may be greater in well
designed facilities, phosphorus removal is more variable. Some cases of negative removal for
ammonia and orthophosphorous are reported; the addition of ammonia or orthophosphorous may
be due, in part, to wildlife use and populations and vegetation management (WY DEQ, 1999).
According to Strecker et al., (1990), overall performance is greatest during the growing season
and lowest during the winter months.
Positive factors influencing pollutant removal include:
constant pool elevations,
range of micro-topography within the watershed,
sediment forebay,
high surface area to volume ratio,
constructed wetland performs better than natural wetland,
adding greater retention volume and/or detention time to the wetland,
effective in areas with high water table or poorly drained soils, and
lengthy travel paths for stormwater.
<
<
<
<
<
<
<
<
Negative factors influencing pollutant removal include::
<
lower removal rate during non-growing seasons,
<
concentrated inflows,
<
sparse wetland cover, and
<
ice cover or snowmelt runoff that would require a modification in design.
3.3.3.3 Design Considerations
General Considerations
A well-designed stormwater wetland consists of:
<
shallow marsh areas of varying depths with wetland vegetation,
<
permanent micropool, and
<
overlying zone in which runoff control volumes are stored.
In addition, all wetland designs must include a sediment forebay at the inflow to the facility to
allow heavier sediments to drop out of suspension before the runoff enters the wetland marsh.
Additional pond design features include an emergency spillway, maintenance access, safety
bench, wetland buffer, and appropriate wetland vegetation and native landscaping (Atlanta
Regional Commission, 2001).
Specific site conditions are important to the proper design of a wetland. Key site characteristics
include soils, hydro period, and plant species and density. Depth to the confining layer or
3 - 31
groundwater is important to ensure that the wetland does not dry up during extended periods of no
rainfall. In addition, a constant source of surface water is recommended, taking appropriate
measures to prevent the undesirable consequences of stagnant water in the wetlands. The depth
and duration of maximum submergence are important because an excess of either will kill the
vegetation (WEF and ASCE, 1998).
Location and Siting
A continuous baseflow or high water table is required to support wetland vegetation. A water
balance must be performed to demonstrate that a stormwater wetland can withstand a 30-day
drought at summer evaporation rates without completely drawing down (Atlanta Regional
Commission, 2001).
Stormwater wetlands should normally have a minimum contributing drainage area of 25 acres or
more and the minimum drainage area is 5 acres for a pocket wetland.
Wetland siting should also take into account the location and use of other site features such as
natural depressions, buffers, and undisturbed natural areas, and should attempt to aesthetically fit
the facility into the landscape. Bedrock close to the surface may prevent excavation.
Stormwater wetlands cannot be located within navigable waters of the U.S., (including wetlands),
without obtaining a Section 404 permit under the Clean Water Act and any other applicable state
permit. In some isolated cases, a wetlands permit may be granted to convert an existing degraded
wetland through local watershed restoration efforts.
Minimum setback requirements for stormwater wetland facilities unless specified by local
ordinances or criteria are as follows:
<
from a property line - 10 ft,
<
from a private well - 100 ft; if well is downgradient from a hotspot land use then the
minimum setback is 250 ft, and
<
from a septic system tank/leach field - 50 ft.
If a wetland facility is not used for overbank flood protection, it should be designed as an offline
system to bypass higher flows rather than passing them through the wetland system.
Sizing
For optimal pollutant removal, a stormwater wetland must meet the following seven basic sizing
criteria:
<
contain a treatment volume (Vt) that is capable of capturing the runoff generated by 70 to
90% of the runoff-producing storms in the region on an annual basis,
<
have a minimum surface area in relation to the contributing watershed area,
<
allocate the surface area of the wetland to meet targets for certain depth zones,
<
meet a minimum standard for the internal flow path through the wetland,
3 - 32
<
<
demonstrate that the water supply to the wetland is greater than the expected loss rate so
that water elevations can be maintained, and
provide for extended detention for smaller storms (for ED wetlands only)(Schueler,
1992b).
Physical Specifications/Geometry
Recommended hydraulic design criteria for wetlands are as follows:
<
Maintain dry-weather flow depths that vary through the wetland between 0.1 and 1.2 m
(0.5 to 4 ft), depending on the types of vegetation planted, with the outlet structure
designed so that the wetland can be periodically drawn down completely to dry the
sediments; this provides for natural oxidation of built-up organics.
<
Size the wet-weather storage volume using the methodology for ED ponds (discussed in
sec 3.3.1.3) but with a maximum surcharge depth above the dry-weather flow depth of 0.6
m (2 ft) and a drawdown time of 24 h; this will reduce stress on herbaceous wetland
plants. The 0.6 m depth limitation will determine the surface area required for the wetland.
<
Design inlet structures to achieve sheet flow across the wetland to the maximum extent
possible.
<
Design the outlet structure to control the water surface and protect it from plugging by
floatables common in wetlands.
<
If open water is to be included in the wetland, it should be less than 50% of the total
wetland area; the depth of the open water should follow the rules for the maximum
permanent pool depth in retention ponds (WEF and ASCE, 1998).
In general, wetland designs are unique for each site and application. However, there are a number
of geometric ratios and limiting depths for the design of a stormwater wetland that are
recommended for adequate pollutant removal, ease of maintenance, and improved safety. Table
3-5 provides the recommended physical specifications and geometry for the various stormwater
wetland design variants. .
3 - 33
Table 3-5 Recommended Design Criteria for Stormwater Wetlands
Design Criteria
Shallow
Wetland
ED Shallow
Wetland
Pond/
Wetland
Pocket
Wetland
Length-to-width ratio
(minimum)
2:1
2:1
2:1
2:1
Extended-detention (ED)
No
Yes
Optional
Optional
Allocation of WQv volume
(pool/marsh/ED) in %
25/75/0
25/25/50
70/30/0
(includes pond
volume)
25/75/0
Allocation of surface area
(deep water/low
marsh/high
marsh/semi-wet) in %
20/35/40/5
10/35/45/10
45/25/25/5
(includes pond
surface area)
10/45/40/5
Forebay
Recommended
Recommended
Recommended
Optional
Micropool
Recommended
Recommended
Recommended
Recommended
Outlet configuration
Reverse-slope
pipe
or hooded broadcrested weir
Reverse-slope
pipe
or hooded
broad-crested
weir
Reverse-slope
pipe
or hooded
broad-crested
weir
Hooded
broadcrested weir
Depth:
Deepwater: 1.5 to 6 ft below normal pool elevation
Low Marsh: 6 to 18 in. below normal pool elevation
High Marsh: 6 in. or less below normal pool elevation
Semi-wet zone: Above normal pool elevation
(Schueler, 1992a; Atlanta Regional Commission, 2001)
Depth
The stormwater wetland should be designed with the recommended proportion of the "depth
zones." Each of the four wetland design variants has depth zone allocations which are given as a
percentage of the stormwater wetland surface area. The four basic depth zones are given below.
Deepwater zone
The deep water zone is between 1.5 and 6 ft deep. It includes the outlet micropool and deepwater
channels through the wetland facility. This zone supports little emergent wetland vegetation, but
may support submerged or floating vegetation.
Low Marsh Zone
Low marsh zone is from 6 to 18 in. below the normal permanent pool or water surface elevation.
This zone is suitable for the growth of several emergent wetland plant species.
3 - 34
High Marsh Zone
This is the zone from 6 in. below the pool to the normal pool elevation. This zone
supports a greater density and diversity of wetland species than the low marsh zone and has a
higher surface area to volume ratio.
Semi-wet Zone
Semi-wet zone refers to those areas above the permanent pool that are inundated during larger
storm events and support a number of species that can survive flooding.
A minimum dry-weather flow path of 2:1 (length-to-width) is required from inflow to outlet
across the stormwater wetland and should ideally be greater than 3:1. This path may be achieved
by constructing internal dikes or berms, using marsh plantings, and by using multiple cells.
Finger dikes are commonly used in surface flow systems to create serpentine configurations and
prevent short-circuiting. Micro topography, or contours along the bottom of a wetland or marsh,
which provides a variety of conditions for different species needs and increases the surface area to
volume ratio, is encouraged to enhance wetland diversity.
A 4 to 6 ft deep micropool should be included in the design at the outlet to prevent the outlet from
clogging and resuspension of sediments, and to mitigate thermal effects. In general, the maximum
depth of any permanent pool areas should not exceed 6 ft.
The volume of the extended detention should not comprise more than 50% of the total WQv, and
its maximum water surface water elevation should not extend more than 3 ft above the normal
pool. Qp and/or Cpv storage can be provided above the maximum WQv elevation within the
wetland.
The perimeter of all deep pool areas (4 ft or greater in depth) should be surrounded by safety and
aquatic benches similar to those for stormwater ponds
The contours of the wetland should be irregular to provide a more natural landscaping effect.
Pretreatment/Inlets
A wetland facility should have a sediment forebay or upstream pretreatment in order to remove
incoming sediment from the stormwater flow prior to dispersal into the wetland. The forebay
should consist of a separate cell, formed by an acceptable barrier and should be provided at each
inlet, unless the inlet provides less than 10% of the total design storm inflow to the wetland
facility.
The forebay should be sized to contain 0.1 in. per impervious acre of contributing drainage and
should be 4 to 6 ft deep. The pretreatment storage volume is part of the total WQv requirement
and may be subtracted from WQv for wetland storage sizing.
A fixed vertical sediment depth marker should be installed in the forebay to measure sediment
3 - 35
deposition over time and the bottom of the forebay may be hardened to make sediment removal
easier.
Inflow channels need to be stabilized with flared riprap aprons, or the equivalent and inlet pipes to
the pond can be partially submerged. Exit velocities from the forebay must be nonerosive.
Outlet Structures
Flow control from a stormwater wetland is typically accomplished with the use of a concrete or
corrugated metal riser and barrel. While the riser is a vertical pipe or inlet structure that is
attached to the base of the micropool with a watertight connection, the outlet barrel is a horizontal
pipe attached to the riser that conveys flow under the embankment. The riser is recommended to
be located within the embankment for maintenance access, safety and aesthetics.
A number of outlets at varying depths in the riser provide internal flow control for routing of the
water quality, channel protection, and overbank flood protection runoff volumes. The number of
orifices varies and is usually a function of the pond design. Alternative hydraulic control methods
to an orifice can be used and include the use of a broad-crested rectangular, V-notch, proportional
weir, or an outlet pipe protected by a hood that extends at least 12 in. below the normal pool.
Higher flows pass through openings or slots protected by trash racks further up on the riser or in a
separate outlet. After entering the riser, flow is conveyed through the barrel and is discharged
downstream. Anti-seep collars should be installed on the outlet barrel to reduce the potential for
pipe failure. Riprap, plunge pools or pads, or other energy dissipators, are to be placed at the
outlet of the barrel to prevent scouring and erosion. If a wetland facility daylights to a channel
with dry weather flow, care should be taken to minimize tree clearing along the downstream
channel, and to re-establish a forested riparian zone in the shortest possible distance.
The wetland facility should have a bottom drain pipe located in the micropool with an adjustable
valve that can completely or partially de-water the wetland within 24 h. However, this
requirement may be waived for coastal areas, where positive drainage is difficult to achieve due to
very low relief.
The wetland drain should be sized one pipe size greater than the calculated design diameter. The
drain valve is typically a handwheel activated knife or gate valve; valve controls shall be located
inside of the riser at a point where they would not normally be inundated, and can be operated in a
safe manner.
Emergency Spillway
An emergency spillway may be included in the stormwater wetland design to safely pass flows
that exceed the design storm flows. It should be located so that downstream structures will not be
affected by spillway discharges. One recommendation is to provide a minimum of 1 ft of
freeboard, measured from the top of the water surface elevation for the extreme flood to the
3 - 36
lowest point of the dam embankment, not counting the emergency spillway.
Maintenance Access
A maintenance right-of-way should be provided to the wetland facility from a public or private
road. Maintenance access should be at least 12 ft wide, have a maximum slope of no more than
15%, and be appropriately stabilized to withstand maintenance equipment and vehicles. The
maintenance access should extend to the forebay, safety bench, riser and outlet, and, to the extent
feasible, be designed to allow vehicles to turn around. Access to the riser should be provided by
lockable manhole covers and manhole steps within easy reach of valves and other controls.
Vegetation
The vegetation diversity in a constructed wetland is established by the landscape plan or volunteer
vegetation. The selection of vegetation should be limited to native plant species suitable for the
pool depths expected within the different depth zones (VA DCR, 1999). Care should be taken to
avoid introducing exotic or invasive species. This problem can be overcome by the use of
appropriate donor soil and wetland mulch. Suitable plants for created wetlands vary between
different eco-regions. However, the wetland plants chosen for created wetlands should
incorporate the following attributes:
<
tolerance to wide ranges of water elevations, salinity, temperature, and pH,
<
a mixture of perennials and annuals,
<
moderate amounts of leaf production, and
<
proven removal efficiencies, e.g., Scriptus spp. (WEF and ASCE, 1998).
Additional considerations include:
<
The use of vegetation and an appropriate landscaping plan should provide elements that
promote greater wildlife and waterfowl use within the wetland and buffers.
<
Woody vegetation may not be planted on the embankment or allowed to grow within 15 ft
of the toe of the embankment and 25 ft from the principal spillway structure.
<
A wetland buffer shall extend 25 ft outward from the maximum water surface elevation,
with an additional 15 ft setback to structures. The wetland buffer should be contiguous
with other buffer areas that are required by existing regulations, or are part of the overall
stormwater management concept plan. No structures shall be located within the buffer and
an additional setback to permanent structures may be provided.
<
Existing trees should be preserved in the buffer area during construction. It is desirable to
locate forest conservation areas adjacent to ponds. Resident goose populations can be
discouraged by planting the buffer with trees, shrubs and native ground covers.
Figure 3-3 shows a schematic of a stormwater wetland (CASQA, 2003).
3 - 37
CIITTAU.
].~. ~m
eYffER L.A!MD:eCAPED
TNE TIlEESiSHRViIS FO!I.~~'T
Lr::1N'1ilAA!".H
_TEIl~Er"fi
~
OM'V)
PLAN VIEW
ANliJ-~~~IJCiOLL4.ft tI'
F"1Ut In'I'tifu'_OOl'
PROFILE
Figure 3-3. Schematic of a Constructed Wetland (CASQA, 2003)
3 - 38
3.3.4 Grassed Swales
The term “grassed swales”, also known as grassed water courses or vegetated swales, refers to the
use of grassed conveyances, which are essentially earthen channels vegetated with erosion
resistant and flood-tolerant grasses. They are designed to infiltrate runoff from intermittent storm
events or to transfer rainfall excess at a non-erosive velocity to desired locations for retention,
detention, storage, or discharge. There are three variations of grassed swales: (1) traditional grass
swales; (2) grass swales with a media filter; and, (3) wet swales. Although using grassed swales
for the sole purpose of conveying stormwater has become a common practice in residential and
institutional settings, their effective use for water quality control is a fairly recent practice, and in
most cases is better accomplished in combination with other BMPs placed downstream to meet
stormwater management requirements (Yousef et al., 1985; MDE, 2000; CASQA, 2003). With
respect to total stormwater management, the desirable attributes in vegetated grass swales include:
<
slower flow velocities than pipe systems, which result in longer times of concentration and
corresponding reduction of peak discharges,
<
ability to disconnect directly connected impervious surfaces, such as driveways and
roadways, thus reducing the computed runoff curve number (CN) and reduction of peak
discharge,
<
filtering of pollutants by grass media,
<
infiltration of runoff into the soil profile, thus reducing peak discharges and providing
additional pollutant removal, and
<
uptake of pollutants by plant roots (phytoremediation) (WEF and ASCE, 1998).
A water quality swale is appropriate where greater pollutant removal efficiency is desired; the
capacity to accept runoff from large design storms being limited in swales, these treatment swales
must often lead into storm drain inlets to prevent large, concentrated flows from gullying/eroding
the swale. Placing check dams across the flow path to temporarily pond runoff could improve the
hydrologic performance of swales with regard to flow attenuation and infiltration for small design
storms (Schueler, 1987).
3.3.4.1 Stormwater Control
Grassed swales and water quality swales usually provide some peak attenuation depending on the
storage volume created by the check dams. However, flood control should be considered a
secondary function of grassed swales since the required storage volume for flood control is
usually more than the swales can provide (NVPDC and ESI, 1992). Swales act to control peak
discharges in two ways:
<
Reduction in runoff velocity by the grass, which depends on the length and slope of the
swale, which in turn, lengthens the time needed for runoff to reach the desired control
point, and can at least partially attenuate the post development peak discharge rate.
<
A portion of the runoff passing through the swale infiltrates into the soil and does not
appear at the downstream control point. However, this seldom exceeds a few tenths of an
inch and depends on soils and slope besides the short contact time of runoff with the swale
3 - 39
(5-20 min.); swale soils have less infiltration capacity than undisturbed soils as they are
heavily compacted to achieve the desired slope and load bearing capacity. Due to previous
saturation of the swale soils by the same rain that supplies runoff to the swale, infiltration
rates in a swale will almost always be near the minimum rates for the local soil type.
3.3.4.2 Pollutant Removal Capability
The primary pollutant removal mechanisms associated with grassed swales are sedimentation and
infiltration into the subsoil. Adsorption and filtration mechanisms can be considered as secondary
removal mechanisms (Schueler, 1987). Changes in the flow hydraulics affected by routing the
flow through grassed channels increase the opportunity for infiltration of soluble pollutants,
deposition of suspended solids, filtration of suspended solids by vegetation, and adsorption of
soluble particles by plants. The flow rate becomes a critical design element since surface runoff
must pass slowly through the filter to provide sufficient contact time for the aforementioned
removal mechanisms to function effectively (NVPDC and ESI, 1992; WY DEQ, 1999).
Conventional grassed swale designs have achieved mixed performance in removing particulate
pollutants such as suspended solids and trace metals. They are generally unable to remove
significant amounts of soluble nutrients. Design practices that increase the retention time of urban
runoff will increase removal efficiencies for soluble forms of nitrogen and phosphorus (Yousef et
al., 1985). Biofilters that increase detention, infiltration, and wetland uptake within the swale
have the potential to substantially improve swale removal rates (CASQA, 2003).
Positive factors influencing pollutant removal (WY DEQ, 1999; CASQA, 2003) include:
<
check dams,
<
low slopes,
<
permeable subsoils and soil moisture holding capacity,
<
dense grass cover or vegetation or mulches,
<
long contact time,
<
smaller storm events,
<
coupling swales with plunge pools,
<
infiltration trenches or pocket wetlands, and
<
swale length greater than one hundred ft.
Negative factors influencing pollutant removal include:
<
compacted subsoils,
<
short runoff contact storms,
<
large storm events,
<
snow melt events,
<
short grass heights,
<
steep slope (6% or greater),
<
runoff velocities greater than 1.5 fps,
<
peak discharge greater than 5 cfs, and
<
dry-weather flow.
3 - 40
3.3.4.3 Design Considerations
The basic design procedure for a swale system was developed by Chow (Chow, 1959) and has
been used in a number of ways in sizing and designing grass swales of varying degrees of
complexity and design robustness. A summary of the various approaches to the design procedure
is presented in this white paper. Reported estimates of low pollutant removal efficiencies for
grassed swales verify the need to improve standard design procedures to make them more
effective for BMP purposes. Studies on swale performance are ambiguous making it hard to
propose specific estimates for swale pollutant removal efficiency. The design of a grassed swale
includes calculations for traditional swale parameters such as flow rate, maximum permissible
velocities, etc. along with storage volume calculations for the water quality volume (VA DCR,
1999). A moderate removal of particulate pollutants can be achieved during small storms if a
swale conforms to the design considerations discussed below.
General Design Recommendations
A dry swale system consists of an open conveyance channel with a filter bed of permeable soils
that overlays an underdrain system. Flow passes into and is detained in the main portion of the
channel where it is filtered through the soil bed. Runoff is collected and conveyed by a perforated
pipe and gravel underdrain system to the outlet. A wet swale consists of an open conveyance
channel that has been excavated to the water table or to poorly drained soils. Check dams are
used to create multiple wetland "cells" that act as miniature shallow marshes. The dry and the wet
swale are designed to treat the WQv through a volume-based design, and to safely pass larger
storm flows. Runoff enters the channel through a pretreatment forebay or along the sides of the
channel as sheet flow through the use of a pea gravel flow spreader trench along the top of the
bank.
Siting
The suitability of a swale at a site will depend on land use, size of the area served, soil type, slope,
imperviousness of the contributing watershed, and dimensions and slope of the swale system
(Schueler et al., 1992). In general, swales (dry/wet) should be sited such that the topography
allows for the design of a channel with sufficiently mild slope and cross-sectional area to maintain
non-erosive velocities (Atlanta Regional Commission, 2001).
Swale siting should also take into account the location and use of other site features such as
buffers, undisturbed natural areas, and natural drainage courses, and should attempt to
aesthetically "fit" the facility into the landscape (CASQA, 2003).
Roadside ditches should be regarded as significant potential swale/buffer strip sites and should be
utilized as such whenever possible. If flow is to be introduced through curb cuts, it is
recommended to place pavement slightly above the elevation of the vegetated areas. Curb cuts
should be at least 12 in. wide to prevent clogging (WEF and ASCE, 1998).
3 - 41
Dry swales can be sited on most soils; however, native soils with low permeability need to be
amended or replaced to increase infiltration. A wet swale can be used where the water table is at
or near the soil surface, or where there is a sufficient water balance in poorly drained soils to
support a wetland plant community (Metro Council, 2001).
The soil below the swale should not consist of too much gravel or coarse sand, as these
constituents do not easily support dense vegetation. It should be undisturbed, as this area may be
periodically inundated and remain wet for long periods of time. In areas with steep slopes, swales
should be employed in locations where they can be parallel to the contours. Unless existing soils
are highly permeable, they should be replaced with a 30 in. depth of a sand/soil mixture
(approximately 50/50 mix) to ensure infiltration.
An underlying engineered soil bed and underdrain system may be utilized in areas where the soils
are not permeable and the swale would remain full of water for extended periods of time (creating
nuisance conditions). This soil bed should consist of a moderately permeable soil material with a
high level of organic matter; e.g., 50% sand, 20% leaf mulch, 30% top soil. The soil bed should
be 30 in. deep and accompanied by a perforated pipe and gravel underdrain system. In residential
developments with marginal soils, it may be appropriate to provide a soil bed and underdrain
system in all grassed swales to avoid possible safety and nuisance concerns.
Appropriate soil stabilization methods, such as mulch, blankets, or mats should be used before
establishing vegetation. Seeding, sodding, and other items related to establishing vegetation
should be in accordance with accepted erosion control and planting practices (WEF and ASCE,
1998; Metro Council, 2001).
Physical Specifications/Geometry
The detention/retention capacity of grassed swales is governed by the runoff associated with the
"water quality storm." The swale length, width, depth, and slope should be designed to
temporarily accommodate the WQv through surface ponding. The WQv is retained for 24 h, but
ponding may continue indefinitely depending on the depth and elevation to the watertable. The
WQv for high density residential, commercial, and industrial land uses will most likely be too high
to be accommodated with most swale designs, and swales in these cases may be appropriate for
pretreatment in association with other practices for these higher density land uses (Metro Council,
2001). The swale can be sized as both a treatment facility for the design storm and as a
conveyance system to pass the peak hydraulic flows of the 100-yr storm if it is located online
(CASQA, 2003).
Water Quality Volume
The purpose of a grassed swale used as a conveyance channel is to transport stormwater to the
discharge point. However, the purpose of a water quality grassed swale is to slow the water as
much as possible to encourage pollutant removal. The use of check dams will create segments of
the swale that will be inundated for a period of time. The required total storage volume behind
3 - 42
the check dams is equal to the water quality volume for the contributing drainage area to that
point. However, the maximum ponding depth behind the check dams should not exceed 18 in.
To insure that this practice does not create nuisance conditions, an analysis of the subsoil is
recommended to verify its permeability.
Swale Geometry
A grassed swale should have a trapezoidal cross-section to spread flows across its flat bottom.
Triangular or parabolic shaped sections are generally not recommended as they tend to
concentrate the runoff. However, a parabolic shape could be acceptable provided the width is
equal to or greater than the design bottom width for a trapezoidal cross section (Atlanta Regional
Commission, 2001; Metro Council, 2001). The side slopes of the swale should be no steeper than
3H:1V to simplify maintenance and help prevent erosion (Schueler, 1987; WA DOE, 2001).
Bottom Width
The bottom width of the swale should be 2 ft minimum and 8 ft maximum in order to maintain
sheet flow across the bottom and avoid concentration of low flows. The 2 ft requirement would
allow for construction considerations and would ensure a minimum filtering surface for water
quality treatment, while the 8 ft maximum would reduce the likelihood of flow channelization
within a portion of the bottom of the swale. Widths up to 16 ft may be used if separated by a
dividing berm or structure to avoid braiding (WEF and ASCE, 1998; VA DCR, 1999; Metro
Council, 2001). The actual design width of the swale is determined by the maximum desirable
flow depth, as discussed below.
Flow Depth
The flow depth for a water quality grassed swale should be approximately the same as the height
of the grass. An average grass height for most conditions is 4 in. Therefore, the maximum flow
depth for the water quality volume should be 4 in. (CWP, 1996b). According to (WEF and
ASCE, 1998), the maximum depth of flow should not be greater than one-third of the gross or
emergent wetland vegetation height for infrequently mowed swales or not greater than one-half of
the vegetation height for regularly mowed swales, up to a maximum of approximately 75 mm (3
in.) for grass and approximately 50 mm (2 in.) below the normal height of the shortest wetland
plant species in the biofilter.
Flow Velocity
The maximum velocity of the water quality volume through the grassed swale should be no
greater than 1.5 fps. The maximum design velocity of the larger storms should be kept low
enough so as to avoid resuspension of deposited sediments. The 2-yr storm recommended
maximum design velocity is 4 fps and the 10-yr storm recommended maximum design velocity is
7 fps.
3 - 43
Longitudinal Slope
The slope of the grassed swale should be as flat as possible, while maintaining positive drainage
and uniform flow, to permit the temporary ponding of the WQv within the channel without having
excessively deep water at the downstream end. The minimum constructable slope is between 0.75
and 1.0% and the maximum slope depends upon what is needed to maintain the desired flow
velocities as well as to provide adequate storage for the water quality volume while avoiding
excessively deep water at the downstream end. Generally, a slope of between 1 and 3% is
recommended. (CASQA, 2003) recommends a maximum of 2.5%, (WEF and ASCE, 1998;
Metro Council, 2001) recommend 2% and (VA DCR, 1999) recommends 3%. The slope should
never exceed 5%.
Swale Length
Swale length is dependent on the swale geometry and the ability to provide the required storage
for the water quality volume. However, the swale should have a length that provides a minimum
hydraulic residence time of at least 10 minutes (CASQA, 2003)), and regardless of the
recommended detention time, the swale should be not less than 100 ft in length.
Swale Capacity
The capacity of the grassed swale is a combined function of the flow volume (the water quality
volume) and the physical properties of the swale such as longitudinal slope and bottom width.
The depth of flow and velocity for any given set of values can be obtained by using the Manning
equation or channel flow nomographs. The Manning's 'n' value, or roughness coefficient, varies
with the depth of flow and vegetative cover. An n value of 0.15 is considered appropriate for flow
depths of up to 4 in.(equal to the grass height). The n value decreases to a minimum of 0.03 for
grass swales at a depth of approximately 12 in. A grassed swale should have the capacity to
convey the peak flows from the 10-yr design storm without exceeding the maximum permissible
velocities. (It must be noted that a maximum velocity is specified for the 2-yr and 10-yr design
storms to avoid resuspension of deposited sediments and other pollutants and to prevent scour of
the channel bottom and side slopes). The swale should pass the 10-yr flow over the top of the
check dams with 6 in. minimum of freeboard. Alternatively, a bypass structure may be
engineered to divert flows from the larger storm events (runoff greater than the water quality
volume) around the grassed swale. However, when the additional area and associated costs for a
bypass structure and conveyance system are considered, it may be more economical to simply
increase the bottom width of the grassed swale. It should then be designed to carry runoff from
the 10-yr design storm at the required permissible velocity. The Manning equation can be used to
adjust the longitudinal slope and bottom width to achieve the maximum allowable velocity (VA
DCR, 1999). The following criteria are probably most applicable in warm and temperate
non-semi-arid climates and should be met or exceeded during the biofiltration capacity design
event (WEF and ASCE, 1998): "maximized" runoff hydraulic residence time of 5 min or more;
maximum flow velocity less than 0.3 m/s (0.9 ft/sec); Manning's n = 0.20 for routinely mowed
swales, and Manning's n = 0.24 for infrequently mowed swales.
3 - 44
One recommended procedure for designing grassed swales (Claytor and Schueler, 1996) outlines
a series of guidelines as shown below and in Table 3-6:
<
Compute the water quality treatment volume (WQv) for the given land surfaces as required
by the local permitting agency.
<
Identify the required swale bottom width, depth, length and slope necessary to store the
WQv within a shallow ponding depth (18 in. maximum).
<
Compute the WQv drawdown time to ensure that it is less than 24 h.
<
Compute the 2-yr and 10-yr frequency storm event peak discharges.
<
Check the 2-yr velocity for erosive potential (adjust swale geometry, if necessary, and
reevaluate WQv design parameters).
<
Check the 10-yr depth and velocity for capacity (adjust swale geometry, if necessary, and
reevaluate WQv design parameters).
<
Provide minimum freeboard above 10-yr stormwater surface profile (6-in. minimum
recommended).
Table 3-6. Design Criteria for Dry (and Wet) Swale Systems
Parameter
Swale Design Criteria
Pretreatment
volume
.05 in. per impervious acre, at initial flow point
Preferred shape
Trapezoidal or parabolic
Bottom width
2 ft minimum, 8 ft maximum; widths up to 16 ft are allowable if a
dividing berm or structure is used
Side slopes
2:1 maximum, 3:1, or flatter preferred
Longitudinal slope
1.0% to 2.0% without check dams
Sizing criteria
Length, width, depth and slope needed to provide surface storage for
WQv. Outlet structures, when used, should be sized to release WQv over
24 h
Underlying soil bed
Equal to swale width
Dry Swale: Moderately permeable soils, 30 in. deep with gravel/pipe
underdrain system if needed
Wet Swale: Undisturbed soils, no underdrain system
Depth and capacity
Surface storage of WQv with a maximum depth of 18 in. for water
quality treatment (12in. average depth)
Safely convey 2 yr storm with non-erosive velocity (# 4.0 ft/s)
Adequate capacity for 10 yr storm with 6 in. of freeboard
(Claytor and Schueler, 1996)
3 - 45
Pretreatment/Inlets
Inlets to swales must be provided with erosion controls as needed (e.g., riprap, flow spreaders,
energy dissipators, sediment forebays, etc.).
Pretreatment of runoff in both dry and wet swale system is typically provided by a sediment
forebay located at the inlet. The pretreatment volume should be equal to 0.1 in. per impervious
acre. A forebay large enough to accommodate 25% of the water quality volume is created by
installing a check dam, constructed of timber or concrete, between the inlet and the main body of
the swale and/or driveway crossings (Atlanta Regional Commission, 2001; Metro Council, 2001).
The checkdam should overlay a stone base to prevent downstream scour. The area downstream of
the checkdam should be protected from scouring with riprap or channel lining. In the undesirable
event of clogging in the surface soils, a checkdam may also be installed at the downstream end of
the swale, along with an optional pea gravel window to route water to the underdrain.
Enhanced swale systems that receive direct concentrated runoff may have a 6 in. drop to a pea
gravel diaphragm flow spreader at the upstream end of the control. A pea gravel diaphragm and
gentle side slopes should be provided along the top of the channels to provide pretreatment for
lateral sheet flows.
Check dams
Check dams are utilized in swales for two reasons: to increase pollutant removal efficiency and/or
to compensate for steep longitudinal slope. The dams should be installed perpendicular to the
direction of flow and anchored into the slope of the channel. The side slopes of the check dams
should be between 5 and 10 to 1 to facilitate mowing operations. The berm height should not
exceed 0.6 m (2 ft), and water ponded behind the berm should infiltrate into the soils within 24 h
(UDFCD, 2002). Check dams should be spaced so that the toe of the upstream dam is at the same
elevation as the top of the downstream dam. For best performance, check dams should have a
level upper surface rather than the uneven surface of a riprap check dam. Earthen check dams are
also not recommended due to erosion potential and high maintenance effort.
Level Spreaders
Level spreaders are diminutive check dams used to provide a uniform flow distribution across the
swale bottom. The hydraulic design of the swale assumes a uniform distribution, which is
difficult to attain without the aid of level spreading devices. The device, placed at the swale inlet,
may consist of a shallow weir across the channel bottom, a stilling basin, or perforated pipe. A
sediment clean-up area should be provided for ease of maintenance.
Flow Bypass
Flow bypass should be considered for high flow events to avoid erosion and channelization. Flow
bypass also allows diversion of flows during swale maintenance, regrading, and vegetation
3 - 46
establishment. Flow can be bypassed by installing a pipe parallel to the swale and a flow
regulating device inside the inlet structure. High flow bypasses may be of two types: "first-flush"
treatment or design flow treatment. The "first-flush" treatment is based on the principle that storm
event pollutants are more concentrated during the "first-flush." Biofiltration swales can be
designed for treating stormwater only from this initial portion of the storm event, and would
require bypassing stormwater flow around the swale during higher portions of flow. More
typically, swale bypasses are designed to treat the design flow throughout the storm event,
bypassing only the flows in excess of the design flow.
Riprap
Riprap is used as an energy dissipation or erosion control device in grassy swales. Riprap pads,
consisting of 152 to 228 mm (6 to 9 in.) rocks that fit tightly across the bed may be used as an
energy dissipater at the swale inlet and continuing for a distance of 1.5 to 3 m (5 to 10 ft)
downstream. Riprap can also be used to line the swale channel if erosion and/or channelization of
the swale bottom are of concern. Riprap could also be used with check dams as described above.
Outlet Structures
Discharges from grassed swales must be conveyed at non-erosive velocities to either a stream or a
stabilized channel to prevent scour at the outlet of the swale. In dry swales, the underdrain system
should discharge to the storm drainage infrastructure of a stable outfall; in wet swales, outlet
protection must be used at any discharge point to prevent scour and downstream erosion.
Emergency Spillway
Enhanced swales must be adequately designed to safely pass flows that exceed the design storm
flows.
Landscaping and Vegetation
Landscape design should specify proper grass species and wetland plants based on specific site,
soil, and hydric conditions present along the channel. A dense cover of water-tolerant,
erosion-resistant grass or other vegetation must be established. Grasses used in swales should
have the following characteristics:
<
a deep root system to resist scouring,
<
a high stem density, with well-branched top growth,
<
tolerance to flooding,
<
resistance to being flattened by runoff, and
<
an ability to recover growth following inundation.
Recommended grasses include but are not limited to the following: Kentucky-31, tall fescue, reed
canary grass, redtop, rough-stalked blue grass, switch grass, little blue stem, and big blue stem. It
should be noted that these grasses can be mixed.
3 - 47
The selection of an appropriate vegetative lining for a grassed swale is based on several factors
including climate, soils, and topography (VA DCR, 1999). Erosion control matting should be
used to stabilize the soil before seed germination. This protects the swale from erosion during the
germination process. In most cases, the use of sod may be required to provide immediate
stabilization on the swale bottom and/or side slopes (WEF and ASCE, 1998). Figures 3-4 and 3-5
provide a representative typical section including both a cross-section and plan view of a dry and
wet swale respectively (MDE, 2000).
3 - 48
( """""c....... nt.~:"
I
,'nO(l,M)"IJ'E.WEfl
J
r W _"".'
,.~
PLAN VIEW
PROFILE
Figure 3-4 Schematic of a Dry Swale (MDE, 2000)
3 - 49
ADDtTJONAL STORAGE
~
\It
~
~
lfll--ll'
~
-ll'
"V-l#-\St~
.,l:.~~.,l:.~~.l.~<l-l.-.l-l.o1'."","~ -\St -\St
.• "
... " "
.~:.:t=~~t~:;;'f'·
.;f""," -.1<
~
\\
PRETREATMENT
(FOREBAY)
"", ... ~.r- ... ~~
OPTlONAL CHECt( DAMS
~
...
..l'-"'-.1<*-.1<*~"'*
RIPRAP
"V
.r-.;f
. . .r-*
.r-.r--.1<
-.l--.l- ... .,l:..;f* ... ~*
"'-.1<*-.1<~~$~~~**.r-.r-.r-*
'*"
of .r- ol- ... \AlETt.AND ~
of ...
*-.1<-.1<",PLANTINGS
-.1< -.l-*01"-.l--.l- ........ ~~~~~~.r-ol-.r-.r-.,l:..,l:..r-
INFLOW
•., ..__
ol-ol-ol-***ol-************ol-,
....... ***... *****.. . ***~~~.~
"" " " "
rSHOULDER
_
ROACWAY-..
PLAN VIEW
2'-8"
aOTTOM WIDTH
WEnAAD
PLANTINGS
SHOULDER·
ROI\rNtIAY
10 YEAR LEVEL
2 YEAR LEVEL
____-=:=~~Imi!~~~~~~~~~~:":~ __lw:A~":R
TABLE (VARIABLE)
V·NOTCH 1NE1R
PROFILE
Figure 3-5 Schematic of a Wet Swale (MDE, 2000)
3 - 50
3.3.5 Vegetated Filter Strips (VFS)
Filter strips are uniformly graded and densely vegetated sections of land, engineered and designed
to treat runoff and remove pollutants through vegetative and soil filtering, evapotranspiration, and
infiltration. Filter strips are similar to grassed swales in many respects, except that they are
designed to accept only overland sheet flow. The requirement that the runoff from an adjacent
impervious area be evenly distributed across the filter strip is not an easy task due to the strong
tendency of runoff to concentrate and form a channel, reducing the performance efficiency of the
filter strip and in some cases leading to erosion of portions of the filter strip (Schueler, 1987).
VFS are best suited to treat runoff from roads and highways, roof down spouts, very small parking
lots, and pervious surfaces. They can function as the outer zone of a stream buffer, or as a
pretreatment for other structural stormwater controls (WY DEQ, 1999; Atlanta Regional
Commission, 2001). To function properly, a filter strip must be equipped with some level
spreading device, densely vegetated with a mix of erosion resistant plant species that effectively
bind the soil, graded to a uniform, even, and relatively low slope, and at least as long as the
contributing runoff area (Knoxville, 2003). There are two filter strip designs: a simple filter strip
and another design that includes a permeable berm at the bottom. The presence of the berm
increases the contact time with the runoff, thus reducing the overall width of the filter strip
required to treat stormwater runoff. As filter strips are typically an on-line practice, they must be
designed to withstand the full range of storm events without eroding.
3.3.5.1 Stormwater Control
Filter strips do not provide enough storage or infiltration to effectively reduce peak discharges to
pre-development levels for design storms (Schueler, 1987; NVPDC and ESI, 1992). The lowering
of runoff velocities and runoff volume, observed sometimes in VFS, may not be typically
adequate for controlling stream channel erosion or flooding (NVPDC and ESI, 1992). Little
attenuation of peak runoff rates and volumes is observed for larger events, depending on soil
properties, suggesting the practice of following strips with another BMP option that can reduce
flooding and erosion downstream (CASQA, 2003). The increasing use of filter strips as a
pretreatment BMP in integrated stormwater management systems helps lower runoff velocities
and hence the watershed time of concentration, slightly reduce both runoff volumes and watershed
imperviousness, and contribute to groundwater recharge.
3.3.5.2 Pollutant Removal Capability
Both swales and filter strips exhibit similar mechanisms of pollutant removal. Pollutant removal
from filter strips is highly variable and depends primarily on density of vegetation and contact
time for filtration and infiltration and soil moisture absorption capacity (Atlanta Regional
Commission, 2001). The mechanisms include the filtering action of vegetation, deposition in low
velocity areas, or by infiltration into the subsoil. The rate of removal appears to be a function of
the length, slope and soil permeability of the strip, the size of the contributing runoff area, and the
runoff velocity (Schueler, 1987).
3 - 51
Vegetated buffer strips are generally effective in reducing the volume and mass of pollutants in
runoff and when designed properly, tend to provide somewhat better treatment of stormwater
runoff than swales with fewer tendencies for flow concentration and the resulting erosion
(CASQA, 2003). Filter strips can effectively reduce particulate pollutant levels in areas where
runoff velocity is low to moderate; however, the ability to remove soluble pollutants under the
same conditions is highly variable (WY DEQ, 1999). Soluble pollutants in filter strips are
removed by pollutant infiltration into the soil and subsequent uptake by rooted vegetation.
However, the efficiency of soluble pollutant removal may not be high since only a small portion
of the incoming runoff will be infiltrated (Schueler, 1987). Filter strips are effective in removing
particulate pollutants such as sediment, organic matter, and trace metals as observed from results
from small test plots and several modeling studies (Schueler, 1987). They also exhibit good
removal of litter and other floatables as the water depth in these systems is well below the
vegetation height. Forested filter strips appear to have greater pollutant removal capability than
grass filter strips because of greater uptake and long-term retention of nutrients in forest biomass.
Positive factors influencing pollutant removal include:
<
minimum strip width of fifty ft,
<
slope of 5% or less,
<
clay soil or organic matter surface,
<
contributing area of less than 5 acre,
<
grass height of 6 to 12 in., and
<
sheet flow.
Negative factors influencing pollutant removal include:
<
runoff velocity > 2.5 fps, depending on site conditions (Horner, 1988),
<
slopes greater than 15%,
<
hilly terrain, and
<
unmowed filter strips.
3.3.5.3 Design Considerations
General Considerations
VFS have limited feasibility as a water quality control in ultra-urban settings with a high
percentage of impervious area where runoff velocities and peak discharge rates are high and flow
is concentrated. Their use is therefore primarily restricted to low and medium density residential
areas (16 to 21% impervious) where they can accept rooftop runoff and runoff from pervious
areas such as lawns, or as a pre-treatment component for structural BMPs in higher density
developments (WY DEQ, 1999). The retrofit capability is relatively simple if enough land area is
available to adequately service the contributing watershed area, and soil and slope conditions are
favorable.
Filter strips should be constructed outside the natural stream buffer area whenever possible to
maintain a more natural buffer along the streambank (Atlanta Regional Commission, 2001).
3 - 52
Forests and other natural areas should not be destroyed to create a filter strip system, as such areas
may already be functional or may only need to be enhanced to function properly as treatment
systems. Disturbance of native vegetation in buffer areas should be avoided whenever possible
(Metro Council, 2001). Other considerations include:
<
Adequate pollutant removal may not be observed on slopes over 15%; filter strips require
climates that can sustain vegetative cover on a year-round basis; contributing upland area
must be small (1 to 5 acres) so that runoff arrives at the filter strip as overland sheet flow;
use of native vegetation or vegetation appropriate for the local climate is essential to
enhance plant survival (WY DEQ, 1999; CASQA, 2003).
<
Filter strips should not be used on soils that cannot sustain a dense grass cover with high
retardance. It is recommended to choose grasses that can withstand relatively high
velocity flows at the entrances during both dry and wet periods (Atlanta Regional
Commission, 2001).
Siting
The use of buffer strips is limited to gently sloping areas where the vegetative cover is robust and
diffuse and where shallow flow characteristics are possible. Slopes should not exceed 15% or be
less than 1% (CASQA, 2003). The vegetative surface should extend across the full width of the
area being drained. The upstream boundary of the filter should be located contiguous to the
developed area. The following site conditions should be considered when selecting a vegetated
filter strip as a water quality BMP:
Vegetated filter strips should be used with soils having an infiltration rate of 0.52 in./h; (sandy
loam, loamy sand). Soils should be capable of sustaining adequate stands of vegetation with
minimal fertilization (VA DCR, 1999). The ability to remove nutrients from surface runoff
improves where clay soils or organic matter are present (WY DEQ, 1999).
A shallow or seasonally high groundwater table will potentially inhibit infiltration and one
recommendation from the VA DCR (VA DCR, 1999) is to have the lowest elevation in the filter
strip at least 2 ft above the water table. Filter strips should be separated from the ground water by
between 2 and 4 ft to prevent contamination and should not remain wet between storms (CASQA,
2003). Greater removal of soluble pollutants can be achieved where the water table is within 3 ft
of the surface, i.e., within the root zone (WY DEQ, 1999). If the soil permeability and/or depth to
water table are unsuitable for infiltration, the primary function of the filter strip becomes one of
filtering and settling of pollutants. This requires a modified design to allow ponding of the water
quality volume at the downstream end of the filter (VA DCR, 1999). Ponding area may be
created by constructing a small permeable berm using a select soil mixture, with the maximum
ponding depth behind the berm as 1 ft.
A natural area that is designed to serve as a vegetated filter strip should not be used for temporary
sediment control.
3 - 53
Flows in excess of design flow must be ensured to move across or around the strip without
damaging it. Higher flows can be handled by a bypass channel or overflow spillway with
protected channel section (Atlanta Regional Commission, 2001).
Physical Specifications/Geometry
Slope
Filter strips appear to be a minimal design practice because they are basically no more than a
grassed slope. The general requirement is that the slope should not exceed 15% and the slope
should be at least 15 ft long to provide water quality treatment. Minnesota urban small sites BMP
manual (Metro Council, 2001) recommends that filter strip slopes should be no less than 1 or 2%
and no greater than 6%. Greater slopes will encourage concentrated flow and flatter slopes may
result in ponding. Both the top and toe of the slope should be as flat as possible to encourage
sheet flow and prevent erosion. The top of the strip should be installed 2 to 5 in. below the
adjacent pavement, so that vegetation and sediment accumulation at the edge of the strip does not
prevent runoff from entering. The flat cross-slope in filter strips ensures that runoff remains as
sheet flow while filtering through the vegetation (CASQA, 2003).
Length
The filter strip should stretch the entire length of the impervious surface from where stormwater
originates and when adjacent to a natural water body, it should stretch the entire length of the
property of shoreline (Metro Council, 2001). VFS should be long enough to provide filtration and
contact time for water quality treatment. Although smaller lengths for e.g., 15 ft, 25 ft are
recommended in some states, (Atlanta Regional Commission, 2001), the recommended minimum
length can be 50 to 75 ft based on the assumption that runoff changes from sheet flow to shallow
concentrated flow after traveling 150 ft over pervious surfaces and 75 ft over impervious surfaces
(CWP, 1996b); an additional 4 ft for any one percent increase in slope is recommended. Filter
strips should be used to treat small drainage areas. Flow must enter the strip as sheet flow spread
out over the width (longer dimension normal to the direction of flow) of the strip, generally no
deeper than 1 to 2 in. However, the length is normally dictated by design method and the
minimum length of a filter strip is:
(3-7)
Wl"uN=Qlq
where
WfMIN = minimum filter strip width perpendicular to flow (ft) (Atlanta Regional Commission,
2001).
The following section highlights the recommendations and minimum design guidelines for
vegetated filter strips intended to enhance water quality. The responsibility rests with the designer
to decide criteria applicable to each facility with any design modifications as well as to provide for
the long-term functioning of the BMP.
3 - 54
Maximum discharge loading per ft of filter strip width (perpendicular to flow path) is found using
Manning's equation:
q 0.00236 Y 513 Sll'
(3-8)
"
where
q = discharge per ft of width of filter strip (ft3/s/ft);
Y = allowable depth of flow (in.);
S = slope of filter strip (%); and
n = Manning's "n" roughness coefficient.
0.15 for medium grass, 0.25 for dense grass, 0.35 for very dense Bermuda-type grass.
Compliance with the design parameters will result in optimal filter strip performance (NVPDC
and ESI, 1992)
Width
Filter strips must be at least 15 ft wide in the direction of flow in order to be effective, however
greater widths will enhance treatment. The steeper the slope, the wider the strip should be.
NRCS recommends a minimum of 150 ft of filtering buffer between a land disturbance activity
and a water body; depending upon soil types and slopes, it may be even greater (Metro Council,
2001).
The width of the filter strip should generally be equal to the width of the contributing drainage
area and when this is not practical, a level spreader should be used to reduce the flow width to that
of the filter strip. The width of the level spreader will determine the depth of flow and runoff
velocity of the stormwater as it passes over the spreader lip and into the filter strip. While a wide
lip will distribute the flow over a longer level section, it reduces the potential for concentrated
flows across the filter (VA DCR, 1999).
A level spreader should be provided at the upper edge of a filter strip when the width of the
contributing drainage area is greater than that of the filter and may extend across the width of the
filter, leaving only 10 ft open on each end. Many configurations of the level spreader can be used
and include a concrete sill or weir, curb stops, curb and gutter with "saw teeth" cut into it, or a
level trench (12 in wide by 24 in deep), filled with pea gravel or crushed stone; the key is to have
a long, continuous, and level overflow elevation to spread the concentrated flow into sheet flow
upstream of the filter strip (VA DCR, 1999; Atlanta Regional Commission, 2001). An effective
technique is to use a pea gravel diaphragm at the top of the slope (a small trench running along the
top of the filter strip) which serves two purposes: (i) it acts as a pretreatment device settling out
sediment particles before they reach the practice; and (ii) it acts as a level spreader, maintaining
sheet flow as runoff flows over the filter strip.
3 - 55
Pervious Berm
A pervious berm may be installed to force ponding in a vegetated filter strip. A pervious berm of
sand and gravel, or soils meeting USDA sandy loam or loamy sand texture, or any other
moderately permeable soil could be installed at the toe of the slope to enhance the effectiveness of
the filter strip. This could also include outlet pipes flowing through the berm or an overflow weir
to provide an area for temporary shallow ponding and accommodate a portion or all of the water
quality volume.
A pervious berm may be installed to force ponding in a vegetated filter strip. It should be
constructed using a moderately permeable soil such as ASTM ML, SM or SC. Soils meeting
USDA sandy loam or loamy sand texture, with a minimum of 10 to 25 % clay, may also be used.
Additional loam should be used on the berm ± 25% to help support vegetation. An armored
overflow should be provided to allow larger storms to pass without overtopping the berm.
Maximum ponding depth behind a pervious berm is 1 ft (VA DCR, 1999; Metro Council, 2001).
Vegetation
A filter strip should be densely vegetated with a mix of erosion resistant plant species that
effectively bind the soil. The selection of plants should be based on their compatibility with
climate conditions, soils, and topography and their ability to tolerate stresses from pollutants,
variable soil moisture conditions, and ponding fluctuations. A filter strip should have at least two
of the following vegetation types:
<
deep-rooted grasses, ground covers, or vines,
<
deciduous and evergreen shrubs, or
<
under- and over-story trees.
Native plant species should be used if possible. As newly constructed stormwater BMPs will be
fully exposed for several years before the buffer vegetation becomes adequately established,
plants that require full shade, or are susceptible to winter kill or prone to wind damage should be
avoided and plant materials should conform to the American Standard for Nursery Stock (VA
DCR, 1999). A schematic representation of a filter strip is given in Figure 3-6 (Claytor and
Schueler, 1996).
3 - 56
t~
(
I
150' Max.
75" Max.
'~l "~"~"~.i~
~
=~~J
aaaaaaaaaaaaraaa~waaWM"
--
--
I --
~
---.------
--.
Planted With Grass Tolerant
--
to Frequent Inundation
t~
,
Maximum
Ponding umit
.II
"-
P ervious
Matenal
Berm
.-
I~'
II.&.
-
.-
25' Min.
length
11'1/'
Pea Gra\ el
Oiaptlr:lgm
Filter Slnp
1
J_
,
Outlet
Pipes, Spaced
@ 25' Center3
/
/'
PLAN
NTS
Grass Filter Strip Length (25' Min.)
Shallow Pending Limit
Pervious Berm
(Sand/Gravel Mix)
Curt>
Stop
Pnng
Lot
\
-..1I./v-,.
Slope Range
2'
~
2% Min.-6°/o Max.
~D~~UUEU~=
=" =.
~
'-~~~QBU~M~OO~Es
'I
12" x 2."
Pea Gravel
=
-
----·""'·rca~D =1'"
Water Quality
Treatment Volume
Diaph~m
j
2
;::11
=,
"'lUUi:i
Outlet
12- Max.
P~s
PROFILE
Figure 3-6. Schematic of a Vegetated Filter Strip (Claytor and Schueler, 1996)
3 - 57
Stream
Forest
Buffer
3.3.6 Infiltration Trenches
A conventional infiltration trench is a shallow, excavated trench, generally 2- to 10- ft deep, that
has been backfilled with a coarse stone aggregate and lined with filter fabric to create an
underground reservoir for stormwater runoff from a specific design storm. Stormwater runoff
diverted into the trench gradually infiltrates into the surrounding soils from the bottom and sides
of the trench. This infiltration reduces the volume of runoff, removes many pollutants and
provides stream baseflow and groundwater recharge. The design storm for an infiltration trench is
typically a frequent, small storm such as the 1- yr event, that provides treatment for the "first
flush" of stormwater runoff (0.5 in. runoff per acre of impervious surface) or even larger volumes.
The trench can be either an open surface trench or an underground facility (VA DCR, 1999) and
infiltration trenches are typically implemented at the ground surface to intercept overland flows
and stormwater runoff that generally enters the facility at one or more point sources. Primarily
used as water quality BMPs, infiltration trench BMPs can route stormwater runoff into the
aggregate filled storage chamber by two means: dispersed input or concentrated input. In
dispersed input, water enters the top of the trench as overland sheet flow directed over a gently
sloping grassed filter strip and flows to the surface of the storage chamber; concentrated input
transports collected runoff to the storage chamber by means of gutters, curb inlets, and pipes.
Some infiltration trench designs combine stormwater detention and water quality objectives by
storing the entire stormwater volume with the water quality volume committed to infiltration by
slowly releasing the water quality volume through an orifice set at a specified level in the storage
reservoir (NVPDC and ESI, 1992).
Infiltration trenches require pretreatment of stormwater in order to remove as much of the
suspended solids from the runoff as possible before it enters the trench. Also, public education
with respect to street/driveway sediments may be provided in areas where an infiltration trench is
proposed (Metro Council, 2001). Enhanced infiltration trenches have extensive pretreatment
systems to remove sediment and oil. Conventional as well as enhanced trenches require on-site
geotechnical investigations to determine appropriate design and location. Generally suited for
low-to medium-density residential and commercial developments, these facilities can be
incorporated in multi-use areas such as along parking lot perimeters, parking lots, residential
areas, commercial areas, and open space areas. Unlike most BMPs, trenches can easily fit into the
margin, perimeter, or other unused areas of developed sites, making them particularly suitable for
retrofitting into existing developments or in conjunction with other BMPs (VA DCR, 1999).
Infiltration may be a more promising practice in that it tends to reverse the hydrologic
consequences of urban development by reducing peak discharges and increasing baseflow to local
streams.
3.3.6.1 Stormwater Control
The size of the infiltration trench is determined by the volume of runoff controlled and the degree
to which infiltration is used to dispose of runoff. There are three basic trench systems.
3 - 58
<
Full exfiltration system - Runoff exits the stone reservoir by exfiltration through the
underlying subsoil and the exfiltration system provides total peak discharge, volume and
water quality control for all rainfall events less than or equal to the design storm (Schueler,
1987). The stone reservoir must be large enough to accommodate the entire increase in
runoff volume for the design storm, less any runoff volume that is exfiltrated during a
storm. Excess runoff from storms greater than the design storm should be handled by an
emergency overflow channel such as a raised curb located above ground.
<
Partial exfiltration system - In this design, an underground drainage system is installed that
comprises regularly spaced perforated pipes located in shallow depressions to collect the
runoff and direct it to a central outlet, and is generally designed to pass the 2-yr storm.
Runoff from smaller storms will still be exfiltrated before it is collected, thereby providing
significant water quality control. An alternative method may be to place perforated pipes
on the underside near the top of the stone reservoir (NVPDC, 1987) to promote a greater
degree of exfiltration, especially for design storms.
<
Water quality exfiltration system - The storage volume of the stone reservoir is generally
set to handle only the first flush of runoff volume during a storm, which has been variously
defined as 0.5 in. of runoff per contributing impervious acre, 0.5 in. runoff per
contributing total acres, and the volume of runoff produced by a 1 in. storm. Runoff
volumes in excess of the first flush are not treated by the system but instead are conveyed
to a stormwater management facility further downstream. While this system does not
satisfy stormwater storage requirements, it may result in smaller, less costly facilities
downstream (Schueler, 1987) .
3.3.6.2 Pollutant Removal Capability
The pollution removal processes that occur in infiltration systems are more complex than those
occurring in wet ponds and extended detention dry ponds. Target pollutant behavior is governed
by an array of factors including pH, redox potential, clay mineralogy, organic matter, microbial
populations and temperature, as well as the physical characteristics of the soil environment, which
change with depth and lateral distance inside the trench. While infiltration trenches are not really
intended to remove a high level of coarse particulate pollutants, which need to be removed by a
pre-treatment device before they enter the trench, fine particulates and soluble pollutants are
effectively removed after exfiltrating through the trench and into the soil (MWCOG, 1979).
Pollutant removal occurs due to sorption, precipitation, trapping, straining and bacterial
degradation and transformation. It should be noted that the pollutant removal capability of water
quality trenches are somewhat lower than other designs as a significant portion of the annual
runoff volume will bypass a water quality trench, and is not subject to removal by exfiltration.
The pollution removal system of an infiltration system has two separate mechanisms. The
sediment control system needed to maintain the function of the trench removes those pollutants
associated with suspended solids such as adsorbed phosphorus, certain heavy metals and some
3 - 59
exchangeable ions. Upon infiltration into the soil, several chemical and biological processes
attenuate the levels of an array of pollutant species (NVPDC and ESI, 1992).
Infiltration trenches eliminate the discharge of the water quality volume to surface receiving
waters and consequently can be considered to have 100% removal of all pollutants within this
volume. Transport of some of these constituents to groundwater is likely, although the attenuation
in the soil and subsurface layers will be substantial for many constituents (CASQA, 2003). The
greatest sorption of nutrients and metals occurs in soils with a high content of clay and/or organic
matter, with the least sorption observed in sandy soils (U.S. EPA, 1977); the same trend holds true
for bacterial densities.
Positive factors influencing pollutant removal include:
<
bank run or washed aggregate,
<
high organic matter and loam content of subsoil,
<
capture of a large fraction of annual runoff volume,
<
effective pretreatment system, e.g., a sump pit; and
<
pretreatment of sediments; oil; and grease.
Negative factors influencing pollutant removal include:
<
sandy soils,
<
trench clogging,
<
high water table,
<
long de-watering times,
<
design considerations, and
<
infiltration trench design variations.
3.3.6.3 Design Considerations
Trench designs can be distinguished as to whether they are located on the surface or below
ground. Surface trenches accept diffuse runoff (sheet flow) directly from adjacent areas after it
has been filtered through a grass buffer. Underground trenches accept more concentrated runoff
(from pipes and storm drains), but require the installation of special inlets to prevent coarse
sediment and oil/grease from clogging the stone reservoir.
Surface trenches are typically applied in residential areas, where smaller loads of sediment and oil
can effectively be trapped by grass filter strips. As the surface is exposed, these trenches have a
slightly higher risk of clogging than underground trenches, which could be prevented by placing a
permeable filter fabric 6 to 12 in. below the surface of the trench for sediment interception. The
following are a few design variations of the surface trench system.
<
Median strip trench design
<
Parking lot perimeter trench design
<
Swale design
3 - 60
Underground trenches can be applied in a variety of development situations and are particularly
suited to accept concentrated runoff. Pretreatment and the even distribution of concentrated
runoff is an essential requirement in these systems. The top of the trench is protected by a layer of
impermeable geo-textile and is covered by topsoil and planted with grass. Underground trenches
may be more aesthetically pleasing, but may also be more expensive to maintain, and, more so
when the trench is covered by pavement or concrete. These BMPs should only be installed when
strong, enforceable maintenance agreements can be secured from the property owner. Some
design variations include:
<
<
<
<
<
over-sized pipe trench;
underground trench with oil/grit inlet;
under-the-swale design;
dry well design; and,
off-line trench system design.
General Considerations
Infiltration can be a very desirable method of stormwater treatment for land uses that do not
heavily pollute stormwater runoff. It may be used where the subsoil is sufficiently permeable to
provide a reasonable infiltration rate and where the water table is low enough to prevent pollution
of groundwater. Areas containing karst topography may initially appear to have excellent
infiltration, but are not recommended for planning an infiltration trench as they may cause
subsurface collapse and sink-hole formation (VA DCR, 1999; Knoxville, 2003).
Paved areas subject to heavy use by motor vehicles, fueling stations, vehicle maintenance
facilities, and similar areas subject to high hydrocarbon loads should be serviced by a water
quality inlet as an in-line pretreatment to any infiltration structure (NVPDC and ESI, 1992).
Infiltration facilities are prone to high failure rates when designed improperly (Schueler, 1992b).
This makes a strong case for designing and accepting infiltration trench systems on the basis of
actual subsurface analysis and permeability tests rather than using pre-existing information on
soils compiled from an array of data (VA DCR, 1999). Further, site-specific soil bores should be
used to justify the use of infiltration practices. A minimum of one soil boring log is recommended
for every 50 ft of trench length, with a minimum requirement of two soil boring logs for each
proposed trench location (Metro Council, 2001). To identify localized soil conditions, soil boring
should be done at the actual location of the proposed infiltration trench. In general, the following
information should be included in a site-specific subsurface or geotechnical study.
Siting
One of the first steps in siting and designing infiltration treatment facilities is to conduct a
characterization study. Geotechnical investigation data can be used for site characterization.
Some of the key data and issues that need characterization include:
<
surface features characterization,
3 - 61
<
<
<
<
subsurface characterization,
infiltration rate determination,
soil testing, and
infiltration receptor (WA DOE, 2001).
Soil Permeability
The soil types within the subsoil profile which extends a minimum of 3 ft below the bottom of the
facility should be identified to verify the infiltration rate or permeability of the soil. The
infiltration rate, or permeability, measured in in./h, is the rate at which water passes through the
soil profile during saturated conditions, the minimum and maximum of which establish the
suitability of various soil textural classes for infiltration. Each soil texture and the corresponding
hydrologic properties within the soil profile are identified through analysis of a gradation test of
the soil boring material. Soil textures acceptable for use with infiltration systems include those
with infiltration rates between 0.52 in./h and 8.27 in./h (VA DCR, 1999), although Schueler
(Schueler, 1987) recommends a minimum infiltration rate of 0.27 in./h (Table 3-7). This implies
that sites with "D" soils (infiltration rates of less than 0.27 in./h), or any soil with a clay content
greater than 30% (as determined from the SCS soil textural triangle) are not suitable options for
infiltration trenches, nor are soils with a combined silt/clay percentage greater than 40% by weight
that are susceptible to frost-heave. Silt loams and sandy clay loams ("C") soils provide marginal
infiltration rates, and should only be considered for partial exfiltration systems. The stone
subgrade must extend below the frost-line irrespective of the soil type, and is typically 8 to 12 in.
in the Washington DC Metropolitan area. Also, trenches should not be located over fill soils that
form an unstable upgrade and are prone to slope failure.
Under suitable soil conditions, soil cores or trenches to a depth of at least 5 ft below the
anticipated level of the stone reservoir bottom may need to be evaluated for any impermeability in
the soil strata that could impede infiltration. However, the presence of such layers does not
necessarily preclude a trench, as long as the stone reservoir completely penetrates them.
Table 3-7. Soil Limitations for Infiltration Trenches
Soil Texture
Effective
Water
Capacity
(Cw )
in. /in.
Minimum
Infiltration
Rate (f)
(in./h)
SCS soil
group
Maximum Depth of Trench
(in.)
48 h
72 h
*Sand
0.35
8.27
A
992
1489
**Loamy Sand
0.31
2.41
A
290
434
**Sandy Loam
0.25
1.02
B
122
183
**Loam
0.19
0.52
B
62
93
3 - 62
Silt Loam
0.17
0.27
C
32
49
Sandy Clay Loam
0.14
0.17
C
20
31
Clay Loam
0.14
0.09
D
11
16
Silty Clay Loam
0.11
0.06
D
7
11
Sandy Clay
0.09
0.05
D
6
9
Silty Clay
0.09
0.04
D
6
7
Clay
0.08
0.02
D
2
4
* Suitable for infiltration with typical 6' to 8' separation from seasonal high groundwater
** Suitable for infiltration with at least 3' separation from seasonal high groundwater
(Schueler, 1987; VA DCR, 1999; Knoxville, 2003)
Depth to Bedrock, Water Table, or Impermeable Layer
Typically, infiltration facilities are not recommended in areas with a high groundwater table due
to the inability of the soil to adequately filter out pollutants before the stormwater enters the table.
While the general requirement of various states is a distance of 2 to 4 ft, the Washington State
Department of Ecology (WA DOE, 2001) recommends that the base of all infiltration trench
systems shall be $ 5 ft above the seasonal high-water mark, bedrock (hardpan) or other low
permeability layer. A minimum separation of 3 ft may be considered if the groundwater
mounding analysis, volumetric receptor capacity, and the design of the overflow and/or bypass
structures are judged to be adequate to prevent overtopping and meet the site suitability criteria
(WY DEQ, 1999).
Topography
The topographic conditions of a development site represent feasibility factors that should be
examined before designing an infiltration system. These factors include the slope of the land, the
nature of the soil (natural/fill), and the proximity of building foundations and water supply wells.
Infiltration trenches should be located in areas in which the slope does not exceed 20% (5H:1V)
because steeper grade would increase the chance of water seepage from the subgrade to the lower
areas of the site and reduce the volume that infiltrates. The use of infiltration trenches on fill
material is not recommended due to the possibility of creating an unstable upgrade. Fill areas can
be very susceptible to slope failure due to slippage along the interface of the in-situ and fill
material, which could be aggravated if the fill material is allowed to become saturated by using
infiltration practices (VA DCR, 1999).
Setback requirements for infiltration trenches that are required by local regulations, uniform
building code requirements, or state regulations generally include the following:
<
>100 ft from drinking water wells, septic tanks or drain fields, and springs used for public
drinking water supplies. Infiltration trenches up gradient of drinking water supplies and
3 - 63
<
<
<
within 1, 5 and 10-yr time of travel zones must comply with Health Department
requirements,
>20 ft downslope and >/100 ft upslope from building foundations,
>20 ft from a Native Growth Protection and Easement (NGPE), and
>50 ft from the top of slopes > 15% (WA DOE, 2001).
On-site and off-site structural stability due to extended subgrade saturation and/or head loading of
the permeable layer need to be evaluated. This would include studying the potential impacts to
downgradient properties, especially on hills with known side-hill seeps.
Design Criteria
Infiltration trenches are assumed to have rectangular cross-sections. Thus, the infiltration surface
area (trench bottom) can be readily calculated from the trench geometry.
Sizing Procedure
The storage volume required for infiltration facilities designed for water quality enhancement is
determined by the water quality volume, determined by the desired pollutant removal efficiency,
and needs to be calculated using the void ratio of the backfill material that will be placed in it.
The sizing of water quality infiltration BMPs is best approached by applying Darcy's Law, which
assumes that the drain time of the facility is controlled by one-dimensional flow through the
bottom surface (VA DCR, 1999).
(3-9)
Q=jxlxSA
where
Q = rate of exfiltration into soil ft3/s;
f = infiltration rate of the soil in ft/h;
I = hydraulic gradient; and
SA = bottom surface area of facility in ft2.
Infiltration Rate
Infiltration rates for treatment can be determined using either a correlation to grain size
distribution from soil samples, textural analysis, or by in-situ field measurements. Short-term
infiltration rates up to 2.4 in./h represent soils that typically have sufficient treatment properties,
while long-term infiltration rates are used for sizing the trench based on maximum pond level and
drawdown time. Long-term infiltration rates up to 2.0 in./h can also be considered for treatment if
site suitability criteria are met for soil infiltration rate/drawdown time as well as soil physical and
chemical suitability for treatment (WA DOE, 2001).
3 - 64
Historically, infiltration rates have been estimated from soil grain size distribution data using the
USDA textural analysis approach. This involves conducting the grain size distribution test on
soils passing the # 10 sieve (2 mm) (U. S. Standard) to determine the percentages of sand, silt and
clay. The ASTM soil size distribution test procedure (ASTM D422), which considers the full
range of soil particle sizes, is also being used by many laboratories to develop soil size
distribution curves; however, these should not be used in conjunction with the USDA soil textural
triangle (WA DOE, 2001).
The three methods for determining the long-term infiltration rate for sizing the infiltration trench
are:
<
USDA soil textural classification,
<
ASTM gradation testing at full scale infiltration facilities, and
<
in-situ infiltration measurements or pilot infiltration tests (PIT) (WEF and ASCE, 1998).
Over the life of the infiltration facility, the rate of infiltration into the soil, f, may gradually
decrease due to clogging of the surface layer of the soil as a result of siltation and biomass buildup
in the trench. This suggests the need for a safety or a correction factor to be built into the design
of the facility to allow for future clogging, which is a factor of 2 to be applied to the infiltration
rate determined from the soil analysis. The design soil infiltration rate, fd, therefore, is equal to
one-half the actual rate:
~
= 0.51
(3-10)
It must be mentioned that a value of 2 for correction factor is based on the assumption that
homogeneous soils should be used for treatment soil suitability determinations (WEF and ASCE,
1998), although a value between 2 and 4 but never less than 2, could be assigned, depending on
the soil textural classification (WA DOE, 1991). These correction factors consider an average
degree of long-term facility maintenance, TSS reduction through pretreatment, and site variability
in the subsurface conditions that affect homogeneity. However, these correction factors could be
reduced, subject to the approval of the local jurisdiction, under the following conditions for sites
with little soil variability:
<
where there will be a high degree of long-term facility maintenance, and
<
where specific, reliable pretreatment is employed to reduce TSS entering the infiltration
facility.
Correction factors higher than the general recommended values should be considered under the
following situations:
<
difficulty in implementing long-term maintenance,
<
little or no pretreatment, and
<
highly variable or uncertain site conditions (WA DOE, 1991).
3 - 65
Hydraulic Gradient
In areas with a shallow water table or impermeable layer, the hydraulic gradient may have an
impact on the allowable design depth. The hydraulic gradient is given by equation 3-11 (VA
DCR, 1999):
(3-11)
where
I = hydraulic gradient;
h = height of the water column over the infiltrating surface (ft); and
L = distance from the top surface of the BMP to the water table, bedrock, impermeable layer, or
other soil layer of a different infiltration rate (ft).
The hydraulic gradient will be assumed to be equal to one in all infiltration designs since the
gradient approaches unity as the facility drains. Therefore,
I=1
Maximum Ponding or Storage Time and Trench Depth
The minimum and maximum time for the trench to empty the stormwater volume into the soil by
infiltration is based upon balancing optimum pollutant removal and assuring adequate stormwater
management performance. Trenches should be designed in general to provide a detention time of
6 to 72 h. A minimum drainage time of 6 h should be provided to ensure satisfactory pollutant
removal in the infiltration trench (Schueler, 1987). Trenches may be designed to provide
temporary storage of stormwater, yet it should drain prior to the next storm event. The drainage
time will vary by precipitation zone and the maximum drain time for the total design infiltration
volume varies from 24 (WA DOE, 2001) to 72 h (Metro Council, 2001). The Northern Virginia
Planning District Commission (NVPDC and ESI, 1992) recommends that the infiltration trench be
designed with a maximum of 48 h for the water quality volume, 72 h for the total volume, and
with a minimum retention time of 24 h for the water quality volume. According to the
Commonwealth of Virginia Department of Conservation and Recreation (VA DCR, 1999),
following the occurrence of a storm event, all infiltration trenches should be designed with a
maximum drain time, Tmax, of 48 h for the water quality volume. The maximum drain time, along
with the minimum design soil infiltration rate, fd, as verified through a subsurface investigation
and analysis, will dictate the maximum allowable design depth, dmax, of the structure. The
maximum depth for an infiltration trench may be defined as:
d. .. • hr...
v,
(3-12)
3 - 66
where
dmax = maximum allowable depth of the trench, in ft;
fd = design infiltration rate of the trench area soils, in ft/h (fd = 0.5f);
Tmax = maximum allowable drain time (48 h)
Vr = void ratio of the stone reservoir expressed in terms of the percentage of porosity divided by
100 (0.4 typically)
A void ratio of 0.40 is assumed for stone reservoirs using 1.5 to 3.5 in. stone - VDOT No. 1
Coarse-graded Aggregate (VA DCR, 1999).
The minimum surface area of the facility bottom may be defined by equation 3-13 :
(3-13)
where
SAmin = minimum trench bottom surface area, in ft2;
Volwq = water quality volume requirements, in ft3;
fd = design infiltration rate of the trench area soils, in ft/h (fd = 0.5f)
Tmax = maximum allowable drain time = 48 h
The storage volume of the facility is defined as: Lx W x D x V r
(3-14)
Determination of the dimensions of the storage reservoir is made by fitting the length, width and
depth into a configuration that satisfies drain-time and storage volume requirements while keeping
the storage reservoir bottom within the optimum depth for infiltration (NVPDC and ESI, 1992).
A long, narrow trench is less affected by water table mounding and is advisable when the depth to
seasonal high water table or bedrock is within 5 ft of the trench bottom. In order to keep the
trench bottom elevation within the optimum depth in the soil profile, long trenches may need to be
curved parallel to the topographic contour. If greater storage is needed than the design storm
volume requirement, the trench dimensions could be adjusted by the following recommendations,
by order of priority:
<
Increase the length of the trench if the seasonal high water table or bedrock is within 5 ft
of the trench bottom.
<
Increase the width, if the length cannot be increased due to site constraints.
<
It is permissible to increase the depth if the seasonal high water table and bedrock are
known to be at a depth greater than 5 ft below the bottom of the trench, provided that the
new bottom elevation meets the same criteria for optimum depth.
Most infiltration trenches are generally greater than 2 ft in depth, but the frost depth needs to be
considered in shallow design trenches. The bottom of the structure should be 18 in. below the
surface to avoid freezing of the trench bottom surface.
3 - 67
Backfill Material
Backfill material for the trench should be clean aggregate with a maximum diameter of 3.5 in. and
a minimum diameter of 1.5 in., and the aggregate should contain few aggregates smaller than the
selected size. An 8 in. bottom sand layer is required for most of the trenches to promote better
drainage and reduce the risk of soil compaction when the trench is backfilled with stone
(Schueler, 1992b).
Filter Fabric
The sides and bottom of the trench should be lined with geotextile fabric (filter fabric). For an
aggregate surface trench, filter fabric should surround all of the aggregate fill material except the
top one foot (VA DCR, 1999). A separate piece of fabric should be used for the top layer to act as
a failure plan. There can be a layer of non-woven filter fabric 6 to 12 in. below the ground surface
to prevent suspended solids from clogging the majority of the storage media. The filter fabric may
need frequent replacement, depending on the volume of suspended solids transported to the
trench.
The filter fabric material must be compatible with the surrounding soil textures and application
purposes, with the cut width of the filter fabric having sufficient material for a minimum 12 in.
overlap. When overlaps are required between rolls, the upstream roll must lap a minimum of two
ft over the downstream roll to provide a shingled effect. The bottom of the infiltration trench can
be covered with a 6 to 12 in. layer of clean sand in place of filter fabric.
Storage Media
The basic infiltration trench design utilizes stone aggregate in the top of the trench to provide
adequate void space (at least 40%) (Schueler, 1987) for filtering and removing pollutants. The
trench should be filled with clean, washed stone with a diameter of 1.5 to 3 in. Pea gravel could
also be substituted for stone aggregate in the top 0.3 meter (1 ft) of the trench, as it improves
sediment filtering and maximizes pollutant removal in the top of the trench. When these modified
trenches become clogged, they can generally be restored to full performance by removing and
replacing only the pea gravel layer, without replacing the lower stone aggregate layers.
Observation Well
An observation well should be installed for every 50 ft of infiltration trench length. The purpose
of the well is to show how quickly the trench dewaters following a storm, as well as providing a
means of determining when the filter fabric is clogged and requires maintenance. It should be
installed in the center of the structure, flush with the ground elevation of the trench. This can be a
4 to 6 in. diameter PVC pipe, anchored vertically to a foot plate at the bottom of the trench, and
the well should have a lockable above-ground cap (Metro Council, 2001).
3 - 68
Overflow Channel
Although an emergency spillway is not necessary because of the small drainage areas controlled
by an infiltration trench, the overland flow path taken by surface runoff when the trench capacity
is exceeded needs to be evaluated. A non-erosive overflow channel leading to a stabilized water
course should be provided, as necessary, to insure that uncontrolled, erosive, concentrated flow
does not develop.
Pretreatment
Infiltration trenches are susceptible to high failure rates due to clogging from sediments, and
therefore require pretreatment of stormwater in order to remove as much of suspended solids as
possible from the runoff before it enters the trench. Pretreatment such as grit chambers, swales
with check dams, filter strips, or sediment forebays/traps should be a fundamental component of
any BMP system relying on infiltration. Pretreatment facilities should be installed off-line in
order to reduce both the frequency of turbulent flow-through and the associated scour and/or
resuspension of residual material (Schueler, 1992b).
A grass strip or other type of vegetated buffer at least 20 ft wide should be maintained around
trenches that accept surface runoff as sheet flow. The slope of the filter strip should be
approximately 1% along its entire length and 0% across its width. A minimum filter length of 50
ft is desirable for areas receiving high loads of suspended solids.
All trenches with surface inlets should be engineered to capture sediment from the runoff before it
enters the stone reservoir. The design of the trench must include a pretreatment facility design,
complete with maintenance and inspection requirements.
Bypass
A bypass system should be implemented for all infiltration trenches. A bypass flow path should
be incorporated in the design of an infiltration trench to convey high flows around the trench. The
overland flow path of surface runoff exceeding the capacity of the infiltration trench should be
evaluated to preclude erosive concentrated flow. If computed flow velocities do not exceed the
non-erosive threshold, overflow may be accommodated by natural topography.
Groundwater Mounding
Groundwater mounding means the local elevation of the water table as a result of infiltrated
surface water, and calculations may be necessary in cases where slope stability is a concern,
and/or a high water table is encountered. The results from these calculations should be regarded
as an indication of the mounding potential rather than as an accurate representation of the actual
mounding depth. Figure 3-7 is an example schematic of an infiltration trench (MDE, 2000).
3 - 69
CONCOEll'
PARKING LOT
l£VEl.
SPREADER
GRASS
""""""-
(leSS TH.A.N ,"SlOPE)
STILLING
BASIN
BYPASS - - - - - - -
(TO DETENTION fACILITY)
f!~"'f
INFILTRATION
TRENCH
WITH PEA GRAVEl
-<
FILTER LAVER
OVER WASHED
BANK RUN GRAVEL
-<
AOOR£OATE
-<
. tr-~"'f •.
. '-""'f'-"
.~ . tC~ .1/:~
'-""'f'-""J
. t!!~ .,..~
'""JIl':!'", JIl':!
Il!~ . tt:~ .
."1',-••.,.,-..
~ • iC~ .If:~
,- .. .,. 1"""'f
.r.~~~
"",
"I'~
'.'
",,- ••i!f~
"l' 1".,
!"'.... •
>>>>-
PLAN VIEW
RUNOFF FlLTEAS THROUGH GRASS
8lJFFER STRIP (20' MINIMUM>; GRASS
CtWmEL: OR SEDIMENT fORE
Y
M1IIi ~-"'"
n~-
-
•
,
rPEAGAAVEL FILTER LAYER
PROTECnve LAVER OF FILTER FABRIC
TRENCH 3-6 FEET DEEP
FUEDWlTH U ·2.5 INCH DIAMETER
CLEAN STONE
(BANK RUN GRAVE"L PREFERREO)
SAND FILTER fr DEEP
(OR fA8JUC EOUIVALENT)
RUNOFF EXFIl.TRATJ;STHROUGH
UNOISTl.RBEO SU8SOllS WITH A
MINIMUM RATE Of D.SINCHES PER HOUR
PROFILE
Figure 3-7. Schematic of an Infiltration Trench (MDE, 2000)
3 - 70
3.3.7 Porous Pavement
Porous concrete and asphalt pavements are being used as BMPs and replace conventional asphalt
pavement or other hard paving surfaces whereby runoff is diverted through a porous asphalt layer
and into an underground stone reservoir. The stored runoff then gradually infiltrates into the
subsoil. The basic porous pavement system consists of a top layer of porous asphalt concrete
covering a layer of gravel that covers a layer of uniformly sized aggregate, which is placed on top
of the existing soil sub-base (Schueler, 1987). Stormwater penetrates the porous asphalt and is
filtered through the first layer of gravel. The voids in the lower level of large aggregate are filled
with runoff. The stored runoff gradually infiltrates into the underlying soil. A sheet of filter
fabric below the aggregate prohibits the underlying soil from entering and clogging the facility
(NVPDC and ESI, 1992). Provided that the grades, subsoil drainage characteristics, and
groundwater table conditions are suitable for its use, porous pavement can be effectively used to
recharge groundwater supplies and reduce stormwater runoff as well as water pollution from
paved low volume traffic areas (WA DOE, 2001). When properly designed and carefully
installed, porous pavement has load bearing strength, longevity, and maintenance requirements
similar to conventional pavement.
The surface of the pavement is designed to provide adequate strength to accommodate vehicles
while allowing infiltration of surface water and filtration of pollutants. If infiltration into the soil
is not practical, the filtered runoff can be discharged through a sub-base drainage system that
would outfall into a storm sewer system or a natural drainage path. Pollutant filtration is greatly
reduced when the pavement drains into a storm sewer. Several studies have concluded that
porous asphalt pavement is sufficiently strong and able to withstand freeze/thaw cycles such that
it will last as long as conventional pavement. Porous pavement systems are typically used in
low-traffic areas such as the following types of applications:
<
parking pads in parking lots,
<
overflow parking areas,
<
residential street parking lanes,
<
recreational trails,
<
golf cart and pedestrian paths, and
<
emergency vehicle and fire access lanes.
There are three types of porous pavement: porous asphalt pavement; porous concrete pavement;
and, modular porous concrete block (WEF and ASCE, 1998).
Porous asphalt pavement is an open-graded coarse aggregate, bound together by asphalt cement
into a coherent mass, with sufficient interconnected voids to provide a high rate of permeability to
water.
Porous concrete (also referred to as enhanced porosity concrete, porous concrete, Portland cement
and pervious pavement) is a subset of a broader family, including porous asphalt, and various
kinds of grids and paver systems. Also known as “no fines concrete,” it is a special type of
concrete that allows stormwater to pass through it, thereby reducing the runoff from a site. In
3 - 71
addition, porous concrete provides runoff treatment through filtration and allows for ground water
recharge. Porous concrete or “no fines concrete paving” is a structural, open textured pervious
concrete paving surface consisting of standard Portland cement, fly ash, locally available open
graded coarse aggregate, admixtures, fibers, and potable water. When properly handled and
installed, porous concrete has a high percentage of void space (approximately 17 to 22%) which
allows rapid percolation of stormwater through the pavement. Porous concrete is thought to have
a greater ability than porous asphalt to maintain its porosity in hot weather and thus is provided as
a limited application control. Although, porous concrete has seen growing use in Georgia, there is
still very limited practical experience with this measure. Porous concrete is designed primarily for
stormwater quality, i.e., the removal of stormwater pollutants. However, they can provide limited
runoff quantity control, particularly for smaller storm events. For some smaller sites, trenches can
be designed to capture and infiltrate the channel protection volume (Cpv) in addition to WQv.
Porous concrete will need to be used in conjunction with another structural control to provide
overbank and extreme flood protection, if required (Atlanta Regional Commission, 2001).
Modular porous pavers are structural units, such as concrete blocks, bricks, or reinforced plastic
mats, with regularly inter-dispersed void areas used to create a load bearing pavement surface.
The void areas are filled with pervious materials (gravel, sand, or grass turf) to create a system
that allows for the infiltration of stormwater runoff. Porous paver systems provide water quality
benefits in addition to groundwater recharge and a reduction in stormwater volume. The use of
porous paver systems results in a reduction of the effective impervious area on a site. There are
many different types of modular porous pavers available from different manufacturers, including
both pre-cast and mold in-place concrete blocks, concrete grids, interlocking bricks, and plastic
mats with hollow rings or hexagonal cells. Modular porous pavers are typically placed on a
gravel (stone aggregate) base course. Runoff infiltrates through the porous paver surface into the
gravel base course, which acts as a storage reservoir as it exfiltrates to the underlying soil. The
infiltration rate of the soils in the subgrade must be adequate to support drawdown of the entire
runoff capture volume within 24 to 48 h. Special care must be taken during construction to avoid
undue compaction of the underlying soils, which could affect the soils’ infiltration capability.
The construction of porous asphalt and concrete are similar to a conventional pavement, except
that sand and finer fraction of the aggregate are left out of the pavement mix, and is typically
placed on top of a granular base. The modular block pavement is constructed by placing the
blocks over a layer of coarse gravel, which in turn is located on a porous geotextile fabric layer.
Porous concrete and asphalt pavements have a tendency to clog and seal within 1 to 3 yr (Urbonas
and Stahre, 1993), with faster sealing rates reported in areas with excessive winter salting and
sanding. Notable exceptions to this were the concrete pavement installations in the state of
Florida. Interlocking cellular concrete block pavement seems to seal at a slower rate and has a
good record of service under a wide range of climatic conditions (WEF and ASCE, 1998).
3 - 72
3.3.7.1 Stormwater Control
Based on the runoff storage provided by the stone reservoir and the degree of reliance on
exfiltration, porous pavement designs fall into three basic categories: complete exfiltration
systems; partial exfiltration systems; water quality exfiltration systems (discussed in 3.3.6.).
3.3.7.2 Pollutant Removal Capability
Porous pavement systems in operation show high removal rates for sediment, nutrients, organic
matter, and trace metals. The majority of the removal occurs as a result of the exfiltration of
runoff into the subsoil, and subsequent adsorption or straining of pollutants within the subsoil
(WY DEQ, 1999). Mechanisms of removal include adsorption, straining, and microbial
decomposition in the subsoil below the aggregate chamber, and trapping of particulate matter
within the aggregate chamber. The first pollutant removal process occurs in the large aggregate
reservoir wherein pollutants adsorb to and are absorbed by the aggregate material. Suspended
matter will settle out at the bottom of the aggregate layer. The second process for removing
pollutants occurs only if the runoff drains into the soil instead of being discharged by a drain.
Pollutants that enter the soil sub-base are also adsorbed to and absorbed by the soil particles in
addition to aerobic decomposition as well as chemical precipitation of the pollutants within the
soil strata (NVPDC and ESI, 1992).
Positive factors influencing pollutant removal include:
<
high exfiltration volumes,
<
high surface area,
<
routine vacuum sweeping,
<
maximum drainage time two days,
<
highly permeable soils,
<
clean-washed aggregate,
<
organic matter in subsoils, and
<
pre-treatment of off-site runoff.
Negative factors influencing pollutant removal include:
<
poor construction practices,
<
inadequate surface maintenance,
<
use of sand during snow conditions, and
<
low exfiltration volumes (WY DEQ, 1999).
3.3.7.3 Design Considerations
Siting
A prerequisite in the construction of porous pavement systems is the evaluation of the site for
feasibility to rely on exfiltration to dispose of runoff. The use of porous pavement is highly
constrained, requiring deep and permeable soils, restricted traffic, and suitable adjacent land uses
3 - 73
Use may also be restricted in regions with colder climates, arid regions, or regions with high wind
erosion rates and in areas of sole-source aquifers (WY DEQ, 1999). Pretreatment using filter
strips or vegetated swales for removal of coarse sediments is recommended (Atlanta Regional
Commission, 2001).
As porous pavements cannot withstand the passage of heavy trucks due to a lower tensile strength
than a conventional pavement, these are typically recommended for lightly used satellite parking
areas and access roads.
The design of porous pavement systems should include a seepage analysis. Possible adverse
impacts of seepage from infiltration measures to building foundations, basements, roads, parking
lots, and sloping areas should be addressed. It is recommended that the porous pavement be
located 10 or more ft down gradient of foundation walls, particularly in residential areas (NVPDC
and ESI, 1992).
Porous pavement systems should be located at least 100 ft away from a drinking water well to
minimize the possibility of groundwater contamination, at least 10 ft down-gradient from nearby
building foundations, and at least 100 ft up-gradient.
Porous concrete systems should typically be used in applications where the pavement receives
tributary runoff only from impervious areas. If runoff is coming from adjacent pervious areas, it
is important that those areas be fully stabilized to reduce sediment loads and prevent clogging of
the porous paver surface. Any significant amount of offsite flow should be diverted around the
pavement surface. Limited offsite runoff and all onsite runoff should be filtered before it flows
over the pavement.
To protect groundwater from potential contamination, runoff from designated hotspot land uses or
activities must not be infiltrated. Porous concrete should not be used under the following
conditions:
<
manufacturing and industrial sites, where there is a potential for high concentrations of
soluble pollutants and heavy metals;
<
areas with a high pesticide concentration; and
<
areas with karst geology without adequate geotechnical testing by qualified individuals
and in accordance with local requirements (Atlanta Regional Commission, 2001).
Soils
Porous pavement is not suitable for sites with soil infiltration rates of less than 0.5 in./h (D soils),
or any soils with a clay content greater than 30%. C soils (silt loam and sandy clay loams)
provide marginal infiltration rates, and should probably only be considered for partial exfiltration
systems (Schueler, 1987; Atlanta Regional Commission, 2001). Soils with a combined silt/clay
content of over 40% by weight are susceptible for frost heave, and may not be suited for these
applications. These systems should never be constructed over fill soils, which often form an
unstable upgrade, and are prone to slope failure. Stone subgrade must extend below the frost line
3 - 74
irrespective of soil conditions. During construction and preparation of the subgrade, special care
must be taken to avoid compaction of the soils.
The most critical factor in determining the applicability of porous pavement as a BMP device is
the infiltration capacity of the underlying soil (NVPDC and ESI, 1992). Core samples or trenches
at least 2 to 4 ft below the anticipated level of the bottom of the stone reservoir should be
examined for any impermeable soil strata that might impede infiltration, such as localized clay
lenses, hardpans, or fragipans. Subsurface drainage may be required if the soil does not exhibit
adequate infiltration capacity. Subsoils are generally susceptible to frost heave if the soil contains
more than 3% of particles smaller than 0.02 mm in diameter. Such soils do not allow the
infiltration from the facility and should be avoided (NVPDC and ESI, 1992).
Slope
Porous concrete systems should not be used on slopes greater than 5%; 2% grade is
recommended. For slopes greater than 1%, barriers perpendicular to the direction of drainage
should be installed in sub-grade material to keep it from washing away, or filter fabric should be
placed at the bottom and sides of the aggregate to keep soil from migrating into the aggregate and
reducing porosity.
Depth to Bedrock and Seasonally High Water Table
The depth from the bottom of the gravel base course to the level of the seasonally high water table
or to bedrock must be sufficient (2 to 4 ft) to allow for adequate infiltration and filtering of water
released through the bottom of the structure. A minimum of 3 ft (preferably 4 ft) of clearance is
needed between the bottom of the stone reservoir and the bedrock level. This data can be inferred
from local soil data maps, but needs to be confirmed by actual soil test bores (WY DEQ, 1999).
To insure complete draining of the stone reservoir in 72 h, it may be necessary to limit the depth
of the stone reservoir if underlying soils have relatively low exfiltration rates. Soil limitations for
porous pavement are shown in Table 3-8.
Table 3-8. Soil Limitations for Porous Pavement
Soil Type
Minimum
Infiltration
Rate (f)
(in./h)
SCS Soil
Group*
Maximum Depth of Storage**
(in.)
48 h
72 h
Sand
8.27
A
992
595
Loamy Sand
2.41
A
290
174
Sandy Loam
1.02
B
122
183
Loam
0.52
B
62
93
3 - 75
Silt Loam
0.27
C
32
49
* Sandy C lay Loams, C lay Loams, S ilty Clay Loams, Sand y Clay, Silty Clay, and Clay soils are not includ ed as these
soil types are all not feasible for infiltration basins.
**M aximu m D epth o f stone re servo ir that can drain c omp letely within 4 8 or 72 h after a sto rm, give n the so il
infiltration ra te .
(Schueler, 1987)
Watershed Size
The most suitable drainage area for porous pavement sites should be restricted to between 0.25
and 10 acres. This guideline tends to reflect the perceived economic and liability problems
associated with larger applications and the cost-effectiveness of other BMPs outside of this range.
Design Parameters
Since the surface area of the porous pavement will typically depend on how large a parking lot
will be built, the critical design consideration will be the depth of the large aggregate layer. As
with infiltration trenches, the maximum depth of the large aggregate layer is a function of
allowable detention time, the porosity of the aggregate, and the soil infiltration rate. The bottom
of the facility should be below the frost line and approximately 4 ft above bedrock and the level of
the seasonally high water table. The same design steps presented for infiltration facilities earlier
in this chapter apply to porous pavement, and include the additional step of determining the
thickness of the porous pavement layer. The depth of the asphalt layer and underlying stone
reservoir depends on the strength of the sub-base soil and the projected traffic intensities (NVPDC
and ESI, 1992).
The following list of general design elements should be considered in any porous pavement
design:
<
anticipated traffic intensities, defined by the average daily equivalent axle load (EAL);
<
California Bearing Ratio (CBR) of the soils; and
<
susceptibility of the soils to forest heave.
Methods for conducting the CBR test are described in ASTM D1883 and AASH0 T193 (VA
DCR, 1999). The asphalt layer is typically 2.5 to 4 in. thick. The minimum combined thickness
of the asphalt layer and stone reservoir can be determined from the Table 3-9.
Table 3-9. Minimum Thickness of Porous Paving
Traffic
Group
General Character
1
Light Traffic
California Bearing Ratio
>15
10-14
6-9
5"
7"
3 - 76
9"
EAL
<5*
<5
2
Medium Light Traffic
(Max. 1,000 VPD)
6"
8"
11"
6-20
3
Medium Traffic
(Max. 3,000 VPD)
7"
9"
12"
21-75
*Studies indicate that for all traffic groups with CBR of 5 or less, the subgrade was improved to CBR 6 with crushed
stone 2 in. size.
VPD = Vehicles Per Day
EA L = E quiva lent Ax le Lo ad (1 8 K ips) av erage daily
Note: Thicknesses refer to the minimum combined d epth of asphalt layer and stone reservoir necessary to carry
appropriate load .
(NVPDC and ESI, 1992)
The following design procedure represents a generic list of the steps typically required for the
design of porous pavement:
<
Determine if the anticipated development conditions and drainage area are appropriate for
a porous pavement application.
<
Determine if the soils (permeability, bedrock, water table, Karst, etc.) and site topographic
conditions (slopes, etc.) are appropriate for a porous pavement application.
<
Locate the porous pavement section on a site within topographic constraints.
<
Determine the drainage area for the porous pavement and calculate the required water
quality volume.
<
Evaluate the hydrology of the contributing drainage area to determine peak rates of runoff.
<
Design the porous pavement stone reservoir; e.g., as shown in the Virginia BMP manual.
<
Design infiltration rate, fd = 0.5 f
<
Max. storage time Tmax = 48 h
<
Max. storage depth, dmax
<
Stone backfill of clean aggregate (1.5 to 3.5 in.) VDOT No. 1 open-graded coarse
aggregate
<
Filter gravel layer - 2 in. of clean aggregate (0.5 in.) VDOT No. 57 open-graded
coarse aggregate
<
Sand layer on trench bottom (8 in.) or filter fabric, per geotechnical and pavement
design recommendations
<
Filter fabric on trench sides and top (not on trench bottom) keyed into trench
<
Overflow channel or large storm bypass.
<
Observation well.
<
Provide pavement section design and material specifications.
<
Provide sequence of construction.
<
Provide maintenance and inspection requirements (VA DCR, 1999).
A schematic representation of a porous pavement is presented in Figure 3-8 (Schueler, 1987).
3 - 77
Sile POS1ed to prev_
'"lH'facing al'ld
Be"" keeps off.site
rvnolf and 5e(Iimen1
OUI. Provides
1~l'npo(,ry Slorage.
Overflow
Ilip e -
Fille,labric .
~
UN
of
abrnl¥fIS Ind 10
resllid tnod< palking
ASphall is vilCWl11 swept
folloWed by jet hosing
~
10 keep pores!"~;;:;.r."",
r;
~~~~~
,Porousasphall
Observation
well
Glilvel
courH 01
Jines mas
of .ow",oi,
to prevent
sedimenlenl/)'
Undistu<bed soils with an fe grealer than 0.27
~sJhout.
6inth
sand layer
preferably O.SO WlthesJhou' 01 more
SldeVteW
Porous pavement COU''Se
(2.5-4.0 lnthes thidt)
• FUte' course
(O.!I inch diameter gravel,
1.0 II1th thidl)
Figure 3-8. Schematic of a Porous Pavement (Schueler, 1987)
3 - 78
3.3.8 Sand and Organic Filters
Sand filters are structural stormwater controls that capture and temporarily store stormwater
runoff and pass it through a filter bed of sand. They have been successfully used in Austin, TX,
the District of Columbia, the state of Delaware, and in Alexandria, VA over the last two decades
(VA DCR, 1999). Most sand filter systems consist of two-chamber structures. The first chamber
is a sediment forebay or sedimentation chamber, which removes floatables and heavy sediments.
The second, which is a filtration chamber, removes additional pollutants by filtering the runoff
through a sand bed. The filtered runoff is typically collected and returned to the conveyance
system, though it can also be partially or fully exfiltrated into the surrounding soil in areas with
porous soils.
Sand filters may be “unconfined” sand-filled trenches with perforated underdrains or “confined”
systems where the filter medium is contained in a concrete vault with a drain at the bottom of the
vault. Depending on the specific design, these types of filters are often referred to as “Delaware
Filters” or “Austin Filters” after the localities where they were originally designed and installed.
Large sand filters are installed above ground and are self-contained sand beds that can treat
stormwater from drainage areas as much as 5 acres in size. Enhanced sand filters utilize layers of
peat, limestone, leaf compost, and/or topsoil, and may also have a grass cover crop. The
adsorptive media of enhanced sand filters is expected to improve removal rates (WY DEQ, 1999).
Sand filters can fall under two basic designs: (i) surface sand filter; and, (ii) perimeter sand filter.
The surface sand filter is a ground-level open air structure that consists of a pretreatment sediment
forebay and a filter bed chamber. This system can treat drainage areas up to 10 acres in size and is
typically located offline. Surface sand filters can be designed as an excavation with earthen
embankments or as a concrete or block structure.
The perimeter sand filter is an enclosed filter system typically constructed just below grade in a
vault along the edge of an impervious area such as a parking lot. The system consists of a
sedimentation chamber and a sand bed filter. Runoff flows into the structure through a series of
inlet grates located along the top of the system.
Yet another design variant, the underground sand filter, is intended primarily for extremely spacelimited and high density areas, and is considered a limited structural application control (Atlanta
Regional Commission, 2001).
3.3.8.1 Stormwater Control
Sand filter systems are designed primarily as offline systems for stormwater quality and typically
need to be used in conjunction with another structural control to provide downstream channel
protection, overbank flood protection, and extreme flood protection, if required. However, under
certain circumstances, filters can provide limited runoff quantity control, particularly for smaller
storm events.
3 - 79
3.3.8.2 Pollutant Removal Capability
Pollutant removal is primarily achieved by straining pollutants through the filtering medium (sand
or peat) and settling on top of the sand bed and/or pretreatment pool. A grass cover crop on the
filter helps in the additional removal of nutrients by plant uptake. Sand filter removal rates are
high for sediment and trace metals, and moderate for nutrients, biochemical oxygen demand
(BOD) and fecal coliform (FC) (City of Austin, 1991).
Positive factors influencing pollutant removal include:
<
offline systems,
<
peat and/or limestone layer,
<
grass cover,
<
longer draw down times ranging from 24 to 40 h,
<
pretreatment pool,
<
minimum depth of 18 in.,
<
regular maintenance, and
<
no direct connection to groundwater.
Negative factors influencing pollutant removal include:
<
online systems, and
<
freezing weather.
3.3.8.3 Design Considerations
General Considerations
Several types of intermittent sand filter facilities are recognized for stormwater quality
management purposes, and the general design criteria presented below apply to the design of these
facilities for water quality control. This implies that the volume of runoff to be treated is
determined by the water quality volume and the desired pollutant removal efficiency (VA DCR,
1999).
The Austin, Texas Filter
The concept and use of surface filters initially originated in Austin, Texas, where these filters have
been extensively used in catchments of up to 20 ha (50 acres). Austin filter has two design
variants, one with full sedimentation and the other with partial sedimentation. The full
sedimentation configuration includes a sedimentation basin designed to hold the entire water
quality volume (i.e., equivalent to the 40 h drain time maximized volume) and to release this
volume to the filter over a 40 h drawdown period. This system should be used unless
topographical constraints make this design unfeasible. The partial sedimentation configuration
requires less depth than the full sedimentation system and may be applicable where topographical
constraints exist. In this system, a smaller sedimentation chamber is located upstream of the
filtration basin, is designed to remove the heavier sediment and trash litter only, and requires more
3 - 80
intensive maintenance than the full sedimentation system. The volume of the sediment chamber
should be no less than 20% of the water quality volume used for the full sedimentation design.
The design must ensure that the sediment chamber discharges the flow evenly(WEF and ASCE,
1998).
Linear Filter - Delaware
The Delaware Filter is an underground system that uses a vault with a permanent pool of water as
the pretreatment device. Recommended for catchments of up to 2 ha (5 acres), the volume of both
the sedimentation and filter chambers are approximated to 38 m3/ha (540 ft3 per contributing acre)
and the surface area of each chamber should be 25 m3/ha. Pavement and inlet design and
construction are critical in a Delaware filter (the filter should be positioned relative to the
pavement to evenly distributed the flow as it enters the sedimentation chamber) (WEF and ASCE,
1998).
Underground Vault - Washington D. C
The initial settling chamber is undersized for effective sedimentation, causing the filter to clog
quickly. When the filter clogs, the flow simply overtops the overflow weir and flows directly to
the outlet, with no indication that the filter is plugged. This filter type when used, should be sized
using the Delaware linear filter criteria, including the pre-settlement chamber. It is also strongly
recommended that the overflow weir and de-watering drain in the filter chamber be blocked and
that the entrance manhole covers over the sedimentation chamber and the outflow chamber be
replaced with grates. If the filter clogs, the water will back up in the vault, overflow out of the
inlet grate over the sedimentation compartment, and back into the outfall chamber, giving a clear
visual indication that the filter is plugged(WEF and ASCE, 1998).
Location and Siting
Surface sand filters should have a contributing drainage area of 10 acres or less. The maximum
drainage area for a perimeter sand filter is 2 acres.
Sand filter systems are generally applied to land uses with a high percentage of impervious
surfaces. Sites with less than 50% imperviousness or high clay/silt sediment loads must not use a
sand filter without adequate pretreatment due to potential clogging and failure of the filter bed.
Any disturbed areas within the sand filter facility drainage area should be identified and stabilized.
Surface sand filters are generally used in an offline configuration where the water quality volume
(WQv) is diverted to the filter facility through the use of a flow diversion structure and flow
splitter. The diversion structure or flow splitter is used to divert stormwater flows greater than the
WQv to other controls or downstream. Perimeter filters are typically sited along the edge or
perimeter of an impervious area such as a parking lot.
3 - 81
Sand filter systems are designed for intermittent flow and must be allowed to drain and re-aerate
between rainfall events and should not be used on sites with a continuous flow from groundwater,
sump pumps, or other sources.
Physical Specifications/Geometry
An access ramp with a slope not exceeding 7:1 or the equivalent should be included at the inlet
and outlet of a surface filter for maintenance purposes. Side slopes for earthen or grass
embankments should not exceed 3:1 (H:V) to facilitate mowing and site slope should be no more
than 6% across filter location (Metro Council, 2001).
A major drawback for a media filtration inlet is the need for elevation differences in the storm
drainage system in order to accommodate live pool storage and sand filter thickness. The
minimum elevation difference needed at a site from the inflow to the outflow is 5 ft for surface
sand filters and 2 to 3 ft for perimeter sand filters (Knoxville, 2003).
A minimum depth of 2 ft is required between the bottom of the sand filter and the elevation of the
seasonally high water table for surface sand filters with exfiltration (i.e,. earthen structure).
While there are no restrictions on the type of soils, Group “A” soils are generally required to allow
exfiltration (Atlanta Regional Commission, 2001).
Sizing
Many guidelines recommend sizing the filter bed using Darcy’s Law, which relates the velocity of
fluids to the hydraulic head and coefficient of permeability of a medium. Hydraulic calculations
based on Darcy’s Law used to establish the filter area of a sand filter allow flow-through of the
treatment volume within the desired time frame, typically 40 to 48 h (City of Austin, 1988; VA
DCR, 1999). The State of Florida uses more complex falling-head computations and allows a
drawdown time of up to 72 h (Florida DER, 1988). However, creating storage for the full WQv in
shallow configuration systems may result in a larger filter than the hydraulic calculations would
indicate (VA DCR, 1999).
The Austin Sand Filter Formula (City of Austin, 1988) derived from Darcy’s Law to size sand
filters is given as:
(3-15)
where
Af = surface area of sand bed (acres or ft2);
Ia = impervious drainage area contributing runoff to the basin (acres or ft2);
H = runoff depth to be treated (ft);
df = sand bed depth (ft);
3 - 82
k = coefficient of permeability for sand filter (ft/h);
h = average depth (ft) of water above surface of sand media between full and empty basin
conditions (½ max. depth) (City of Austin, 1996); and
tf = time required for runoff volume to pass through filter media (h)
A BMP drawdown time (tf) of 40 h allows the filter to fully drain down and dry out to maintain an
aerobic environment between storms. Typical values for k are shown in Table 3-10.
Table 3-10. Coefficient of Permeability k Values for Stormwater Filtering Practices
Filter Medium
Coefficient of Permeability (ft/d)
Sand
3.5
Peat/Sand
2.75
Compost
(CWP, 1996a)
8.7
The permeability of sand shown in Table 3-10 is extremely conservative, but is widely used since
it is incorporated in the design guidelines of the City of Austin(City of Austin, 1988; City of
Austin, 1996). When the sand is initially installed, the permeability is so high (over 100 ft/d) that
generally only a portion of the filter area is required to infiltrate the entire volume, especially in a
“full sedimentation” Austin design where the capture volume is released to the filter basin over 24
h. This methodology results in a filter bed area that is oversized when new and the entire water
quality volume is filtered in less than a day with no significant height of water on top of the sand
bed. The Austin design variations are still preferred where there is sufficient space, because they
lack a permanent pool, which eliminates vector concerns. Consequently, the simple rule of thumb
is adequate for sizing the filter area.
For filters with full sedimentation protection (sedimentation basin containing full WQv with 24 h
drawdown to filter), k = 3.5 ft/d (0.146 ft/h), and tf = 40 h, the sand filter formula reduces to:
Af(n)=
3lOJ.dr
(h+d )
r
(3-16)
where
Af is in ft2 and Ia is in acres.
For filters with partial sedimentation protection (sediment chamber containing 20% of WQV with
free hydraulic flow to filter), k = 2.0 ft/day (0.0833 ft/h) and tf = 40 h, the formula reduces to:
3 - 83
545I~d r
Af(pS}
(h+d
r
)
(3-17)
Where
Af is in ft2 and Ia is in acres.
Capture volume
The facility should be sized to capture the required water quality volume, preferably in a separate
pretreatment sedimentation basin.
Geometry
The water depth in the sedimentation basin when full should be at least 2 ft and no greater than 10
ft. A fixed vertical sediment depth marker should be installed in the sedimentation basin to
indicate when 20% of the basin volume has been lost because of sediment accumulation.
Basic Components
Surface sand filters generally employ the following layers, from top to bottom: sand, geotextile
and an underdrain system. Runoff discharging to the sand filter must be pretreated (e.g., a
presettling basin) to remove debris and other gross solids and any oil from high-use sites. The
type of pretreatment device will depend on the type of pollutants present. The length-to-width
ratio of the presettling basin should be 3:1 and the recommended depth varies from 3 to 6 ft.
Inlet structures such as flow spreaders, weirs or multiple orifice openings should be designed to
minimize turbulence and spread the flow uniformly across the surface of the filter media.
Stone riprap or other dissipation devices should be installed to prevent gouging of the sand media
and promote uniform flow. Offline outlet structures are typically sized for the 15-min peak flow
of a 2-yr, 24-h storm.
An impermeable liner (clay, geomembrane or concrete) may be required under the filter to protect
groundwater or where underflow could damage structures. If the impermeable liner is not
required, a geotextile liner should be installed, unless the bed has been excavated to bedrock
(Metro Council, 2001).
The sand filter is typically constructed with 18 in. of sand overlying 6 in. of gravel. The sand and
gravel media are separated by permeable geotextile fabric and the gravel layer is situated on
geotextile fabric. Four-in.perforated PVC pipe is used to drain captured flows from the gravel
layer. A minimum of 2 in. of gravel must cover the top surface of the PVC pipe.
3 - 84
Sand Specification
The sand in a filter must consist of medium-sized sand that meets the size gradation (Table 3-11).
A laboratory analysis to determine the sand’s hydraulic conductivity K is also highly
recommended. The designer should then adjust this number to account for conditioning of the
sand during operation.
Table 3-11. Sand Medium Specification
U. S. Sieve Number
Percent Passing
4
95-100
8
70-100
16
40-90
30
25-75
50
2-25
100
<4
200
(King County, 1998)
<2
Underdrain Systems
Several types of underdrains may be used: a central collector pipe (with lateral feeder pipes or a
geotextile drain strip in an 8 in. gravel backfill or drain rock bed) or a longitudinal pipe in an 8 in.
gravel backfill or drain rock with a collector pipe at the outlet end.
Hydraulically, the system is typically sized for the 15 min peak flow from a 2-yr, 24-h storm, with
1 ft of head above the invert of the upstream end of the collector pipe. Yet, local sizing
requirements should be used when available.
A geotextile fabric must be used between the sand layer and drain rock or gravel and placed so
that 1.0 in. of drain rock or gravel is above the fabric. Drain rock should be 1.5 to 0.75 in. rock or
gravel backfill, washed free of clay and organic material.
Cleanout wyes with caps or junction boxes must be provided at both ends of the collector pipes.
Cleanouts must extend to the surface of the filter. A valve box must be provided for access to the
cleanouts.
3 - 85
Impermeable Layers
Impermeable liners such as clay, concrete, or geomembrane should be used when nonconventional soluble pollutants such as metals or organics are present, and where the underflow
could cause problems with structures or groundwater. Clay liners should have a minimum
thickness of 12 in. and meet the specifications in Table 3-12.
A geomembrane liner should be at least 30 mils thick and ultraviolet resistant. It should be
protected from puncture, tearing and abrasion by installing geotextile fabric on the top and bottom
of the geomembrane.
Concrete liners may also be used for basins less than 1,000 ft2 in area. Concrete should be 5 in.
thick (Class A or better) and reinforced by steel wire mesh. The concrete should have a minimum
6 in. compacted aggregate base consisting of either coarse sand and river stone or crushed stone or
its equivalent with diameter of 0.75 to 1 in., when the underlying soil is clay or has an unconfined
compressive strength of 0.25 ton/ft2 (Metro Council, 2001).
If an impermeable liner is not provided, an analysis should be made of possible adverse impacts of
seepage zones on groundwater and nearby built areas. Sand filters without impermeable liners
should not be built on fill sites, and should be located at least 20 ft downslope and 100 ft upslope
from building foundations (Metro Council, 2001).
Table 3-12. Clay Liner Specifications
Property
Test Method
Unit
Specification
Permeability
ASTM D-2434
cm/sec
1 × 10 -6 max
Plasticity index of clay
ASTM D-423 & D-424
percent
Not less than 15
Liquid limit of clay
ASTM D-2216
percent
Not less than 30
Clay particles passing
ASTM D-422
percent
Not less than 30
Clay compaction
ASTM D-2216
percent
95% of Standard
Proctor Density
(WA DOE, 2001)
Underground Filters
Although, in general, sand filter design criteria apply to underground filters as well, additional
specific recommendations for underground filters are as follows:
<
One ft of sediment storage in the presettling cell should be provided.
3 - 86
<
The retaining baffle for oil/floatables in the pre-settling cell must extend at least 1 ft above
to 1 ft below the design flow water level, and be spaced a minimum of 5 ft horizontally
from the inlet. Provision for the passage of flows in the event of plugging must be
provided. Access opening and ladder must be provided on both sides of the baffle.
<
The inlet flow distribution should be optimized with minimal sand bed disturbance. One
recommendation is to provide a maximum of 8 in. of distance between the top of the
spreader and the top of the sand bed. Flows may enter the sand bed by spilling over the
top of the wall into a flow spreader pad. Alternatively a pipe and manifold system may be
used. Any pipe and manifold system must retain the required dead storage volume in the
first cell, minimize turbulence, and be readily maintainable. Multiple inlets are
recommended to minimize turbulence and reduce local flow velocities.
<
Erosion protection must be provided along the first foot of the sand bed adjacent to the
spreader. Geotextile fabric secured on the surface of the sand bed, or an equivalent
method may be used. A dewatering gate valve should be constructed just above the sand
bed and removable sand panels must be provided over the entire sand bed.
<
To prevent anoxic conditions, a minimum of 24 ft2. of ventilation grate must be provided
for each 250 ft2 of sand bed surface area. For sufficient distribution of air flow across the
sand bed, grates may be located in one area if the sand filter is small, but placement at
each end is preferred. Small grates may also be dispersed over the entire sand bed area.
Organic Filters
The organic filter is a design variant of the surface sand filter, which uses organic materials such
as leaf compost or a peat/sand mixture as the filter media. The organic material enhances
pollutant removal by providing adsorption of contaminants such as soluble metals, hydrocarbons,
and other organic chemicals (Atlanta Regional Commission, 2001). Additional specific
recommendations for organic filters are as follows: The type of peat used is critically important.
Fibric peat, in which undecomposed fibrous organic material is readily identifiable, is preferred.
Hemic peat containing more decomposed material may also be used. Sapric peat, made up largely
of decomposed matter, is not recommended. They are typically used on relatively small sites (up
to 10 acres), to minimize potential clogging. The minimum head requirement (the elevation
difference needed at a site from the inflow to the outflow), 5 to 8 ft, is higher than the surface sand
filter.
Two typical media bed configurations are the peat/sand and the compost filter. Both variants
utilize a gravel underdrain system. The peat filter includes an 18 in. 50/50 peat/sand mix over a 6
in. sand layer and can be optionally covered by 3 in. of topsoil and vegetation. The compost filter
has an 18 in. compost layer.
Figures 3-9 to 3-12 show the schematic of a surface sand filter, perimeter sand filter, underground
sand filter, and organic filter, respectively.
3 - 87
BYPASS
UNDERORAIN COLLECnON
SYSTEM
now DIVERSION
FILTER RED
i
~
,,••
,
..
--~
QUIfJ,QW ..
... ' .....OVERFLOW
... ~~
SPILLWAY
••
.....
t
~
PRETREATMENT
SEDIMENTloTION
CHAMDER
PLAN VIEW
PERFORATED
STANDPIPE
OVERFLOW srtu'WAY
FlL
/
R BED
•
OllTFLOW
UNMRDRAIN COLLECTION SYSTEM
PROFILE
)"TOPSOII.
"~
~: -~"
'1= 1'1::'
ill "'- ~'':':
IS" CLEAN WASHED
"CONCRETI:" SAND
':I=lI.,::
11;1
I'll·· •
11,1_
l=l _·':1:112 t
J:'fi~:
~~ .C
'"
.:-u·· .:'_. 1:'1
__ II'· I
.:'~::'/'J~~~-..
.:;~q.:..
.':1
~-=n
.oJ J . I I
::. Ii''':. ~
'li
l:I
GEOTEXTII.E
Au' SIDES, TOP &. BOTTOM
I
6" PERFORATED PIPEIGRAVEL
UNDERDRArN SYSTEM
TYPICAL SECTION
Figure 3-9. Schematic of a Surface Sand Filter (MDE, 2000)
3 - 88
pAAJ<1NG lOT SHEET FLOW
INLET GAATES
CURB STOPS
OIIERFlOWWEIRS
'"SO'" %3
'.'
.
'"
,""',. .,..,
.. -
'"
.
CLEAR
","ll
INlET GRATES
ACCESS GAATES
'S7
TEI.lPORAAY PONDING
","'R
SEDlMENTATlOH
CHAM"'R
UNOERDRAIN
.... ;-:: .-:::- ·Iltl=
I=llll-
PROFILE
la'ClEAN
WASHED SAND
Figure 3-10. Schematic of a Perimeter Sand Filter (MDE, 2000)
3 - 89
UNOERORAIH
AOCESSGRATES
""""
PIPE SYSTEM
~p.
"""HOU;
t
. '.
o
RFlOW ."
" SEOII'Il;IUATIONCHA
R
""""'"R
'"
........•. ,.
'.,
'.,
...
'"
.. ,
'"
.. ,'
PLAN VIEW
ACCESS GRATeS ~
_STEPS
(TVp,
TEWPOfW<'
PONOINO
OVERFlON
WEIR
TEMPORARY
PQNOING
(VAR>AalE,
OEBRJS SCREEN (I,
6" PERfORATED PIPE
IN
ll"GAA\/1:l.LACXETJ,-_":;:;;:~~-:.~
TYPICAL SECTION
TYPICAL SECTION
Figure 3-11. Schematic of an Underground Sand Filter (MDE, 2000)
3 - 90
UNOERDRAlN COLLECTION
SYSTEM
ORGANIC FILTER 8£0
FLOW DIVERSION
••
.~
...
OIITfl,OW ..
•• - ....-OVERFLOW
.-',
SPILLWAY
••
....
PRETREATMENT
SEDIMENTATION
CliAMBER
PLAN VIEW
PERFORATED
STANDPIPE
/VERFLOW SPILLWAY
FILTE.R BED
OUTFLOW
UNDERDRAIN COLLECTION SYSTEM
PROFILE
3" TOPSOIL
+-~
3"TOPSOIL
_11~
18" SO/SO
PEAT/SAND
MIXTIlRE
I
... III_,.J_ :':l_llII..:11
-jl!
rJ
lW
1
I~:
18" TO 24"
~illl-lII" ] .. n:!=rlill: II
1_lllI~lIl.=e~_
! Il!i
'::'1lI_HI_
~
,1:'*,_"
'"
,~,
."., ".... "
'
..
"
'
__ l!1
fa
, "
LEAF
COMPOST
=n
I_fli'
., " ,.~
" ., . . ,.
GEOTEXT1LE
..
ALL SIDES.
,~.
,
..
."
"
,,'
-.'0'
l'
,.'
"
.~._
-"
d' :\' 'I
,p' .'"
.,
.j.
."
,('.,
"
•• ,
"
. '.
..1
.Ii· ",
:1'
,,' ",
.'\
,,- .'1' .j' .' .,l ,j>
• "' > -{\
:,', , • .",
• ~"
TOP&:. BOTTOM
"
d'
·'_-.",~'.l;···"."}::·,.i"
6- PERFORATED
PIPE IGRA VEL
UNDERDRAIN SYSTEM
TYPICAL SECTIONS
Figure 3-12. Schematic of an Organic Filter (MDE, 2000)
3 - 91
REFERENCES
Atlanta Regional Commission. (2001). Georgia Stormwater Management Manual Volume 2:
Technical Handbook. Atlanta Regional Commission, www.georgiastormwater.com
California Department of Transportation (Caltrans). (2002). Proposed Final Report: BMP Retrofit
Pilot Program. CTSW-RT-01-050.
California Stormwater Quality Association (CASQA). (2003). Stormwater Best Management
Practice Handbook: New Development and Redevelopment. www.cabmphandbooks.com.
Chow, V.T. (1959). Open-Channel Hydraulics. McGraw Hill, New York, NY.
(City of Austin). (1988). Environmental Criteria Manual. Austin, TX.
(City of Austin). (1991). Water Quality Management. Environmental Criteria Manual. Austin,
TX.
(City of Austin). (1996). Design of Water Quality Controls. Austin, TX.
Clar, M., B.J. Barfield, and S. Yu. (2003). Considerations in the design of treatment best
management practices (BMPs) to improve water quality. EPA 600/R-03/103. U.S.
Environmental Protection Agency, Office of Research and Development, Cincinnati, OH.
www.epa.gov/ORD/NRMRL/Pubs/600r03103/600r03103.htm October 2003.
Claytor, R.A. and T.R. Schueler. (1996). Design of Stormwater Filtering Systems. Center for
Watershed Protection and Chesapeake Research Consortium, Ellicott City and Solomons,
MD.
Center for Watershed Protection (CWP). (1996a). Design of Stormwater Filtering Systems.
Ellicott City, MD.
Center for Watershed Protection (CWP). (1996b). Ditches or Biological Filters? Classifying the
pollutant removal performance of open channels. In Watershed Protection Techniques,
Volume 2 No. 2. Ellicott City, MD.
Florida DER. (1988). The Florida Development Manual: A Guide to Sound Land and Water
Management. Florida Department of Environmental Regulation, Tallahassee, FL.
Guo, C.Y. and B.R. Urbonas. (1995). Special Report to the Urban Drainage and Flood Control
District on Stormwater BMP Capture Volume Probabilities in United States,. Denver, CO.
Horner, R.R. (1988). Biofiltration Systems for Storm Runoff Water Quality Control. Washington
State Department of Ecology, Seattle, WA.
King County Department of Natural Resources (King County). (1998). King County Washington
Surface Water Design Manual. Seattle, WA.
City of Knoxville (Knoxville). (2003). Best Management Practices (BMP) Manual. Knoxville,
TN.
Maryland Department of the Environment (MDE). (2000). Maryland Stormwater Design Manual
Volumes I & II. Baltimore, MD. http://www.mde.state.md.us.
(Metro Council). (2001). Minnesota Urban Small Sites BMP Manual. Stormwater Best
Management Practices for Cold Climates. St. Paul, MN.
Metropolitan Washington Council of Governments (MWCOG). (1979). Controlling Stormwater
Runoff in Developing Areas: Selected Best Management Practices. Washington, DC.
Northern Virginia Planning District Commission (NVPDC). (1979). Guidebook for Screening
Urban Nonpoint Pollution Management Strategies. Annandale, VA.
3 - 92
Northern Virginia Planning District Commission (NVPDC). (1987). BMP Handbook for the
Occoquan Basin (Draft). Annandale, VA.
Northern Virginia Planning District Commission and Engineers and Surveyors Institute (NVPDC
and ESI). (1992). Northern Virginia BMP Handbook: A guide to planning and designing
best management practices in northern Virginia. Annandale, VA.
Schueler, T.R. (1987). Controlling Urban Runoff: A practical manual for planning and designing
urban BMPs. Metropolitan Washington Council of Governments, Washington, DC.
Schueler, T.R. (1992a). Design of Stormwater Wetland Systems: guidelines for creating diverse
and effective stormwater wetlands in the mid-Atlantic Region. Metropolitan Washington
Council of Governments, Washington, DC.
Schueler, T.R. (1992b). Design of Stormwater Pond Systems. Metropolitan Washington Council
of Governments,, Washington, DC.
Schueler, T.R., P.A. Kumble, and M.A. Hearty. (1992). A Current Assessment of Urban Best
Management Practices: Techniques for reducing nonpoint source pollution in the coastal
zone. Metropolitan Washington Council of Governments, Washington, DC.
Strecker, W.E., J.M. Kersnar, and E.D. Driscoll. (1990). The use of wetlands for controlling
stormwater pollution. Woodward-Clyde Consultants. Prepared for the United States
Environmental Agency.
United States Environmental Protection Agency (U.S. EPA). (1977). Process Design Manual for
Land Treatment of Municipal Wastewater. EPA-625-1-77-008. Washington, DC.
United States Environmental Protection Agency (U.S. EPA). (1983). Results of the Nationwide
Urban Runoff Program. Volume I - Final Report. EPA-832-R-83-112. Washington, DC.
United States Environmental Protection Agency (U.S. EPA). (1996). Overview of the Stormwater
Program. EPA-833-R-96-008. Washington, DC.
Urban Drainage and Flood Control District (UDFCD). (2002). Urban Storm Drainage Criteria
Manual: volume 3 - best management practices. Denver, CO. www.udfcd.org.
Urbonas, B.R. and P. Stahre. (1993). Stormwater - Best Management Practices Including
Detention. Prentice Hall, Englewood Cliffs, NJ.
Commonwealth of Virginia Department of Conservation and Recreation (VA DCR). (1999).
Virginia Stormwater Management Handbook, Volume I,. Richmond, VA.
Washington State Department of Ecology (WA DOE). (1991). Stormwater Management Manual
for the Puget Sound Basin - Public Review Draft. Olympia, WA.
Washington State Department of Ecology (WA DOE). (2001). Stormwater Management Manual
for Western Washington. Volume V - Runoff Treatment BMPs.
http://www.ecy.wa.gov/programs/wq/stormwater/index.html.
Water Environment Federation and American Society of Civil Engineers (WEF and ASCE).
(1998). Urban runoff quality management. 1-57278-039-8 and 0-7844-0174-8.
Alexandria, VA.
Wyoming Department of Environmental Quality (WY DEQ). (1999). Urban Best Management
Practices for Nonpoint source Pollution.
Yousef, Y., M. Wanielista, and H. Harper. (1985). Best Management Practices: Removal of
highway contaminants by roadside swales. Prepared for Florida Department of Highway
Transportation, Tallahassee, FL.
3 - 93
4
BMP Monitoring
Bethany Madge
CONTENTS
4.0 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 2
4.1 Current Monitoring Practices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 3
4.2 Parameter Selection for Structural BMP Monitoring Programs . . . . . . . . . . . . . . . . . . . . 4 - 5
4.2.1 Major Parameter Categories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 6
4.2.1.1 Chemical Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 6
4.2.1.2 Physical Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 7
4.2.1.3 Biological Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 8
4.2.1.4 Hydrologic and Hydraulic Parameters . . . . . . . . . . . . . . . . . . . . . . . . 4 - 10
4.2.1.5 Contributing Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 11
4.2.2 Key Considerations for Selecting Appropriate Parameters . . . . . . . . . . . . . . . . 4 - 12
4.3 Monitoring Nonstructural BMPs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 14
4.4 Watershed Monitoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 15
4.5 Developing a BMP Effectiveness Monitoring Program . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 17
4.5.1 The Planning Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 17
4.5.1.1 Defining Program Goals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 17
4.5.1.2 Collecting Background Information . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 18
4.5.1.3 Identifying Project Resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 18
4.5.1.4 Formulating Monitoring Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 19
4.5.2 The Design Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 19
4.5.2.1 Monitoring Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 19
4.5.2.2 Parameter and Methods Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 20
4.5.2.3 Hydrologic and Hydraulic Data Collection . . . . . . . . . . . . . . . . . . . . 4 - 20
4.5.2.4 Water Quality Data Collection Protocols . . . . . . . . . . . . . . . . . . . . . . 4 - 22
4.5.2.5 Selection of Equipment and Materials . . . . . . . . . . . . . . . . . . . . . . . . 4 - 26
4.5.2.6 QA/QC Initiatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 26
4.5.2.7 Quality Assurance Project Plan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 27
4.5.3 The Implementation Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 27
4.5.4 The Evaluation Phase - Quantifying BMP Efficiency . . . . . . . . . . . . . . . . . . . 4 - 27
4.5.5 The Presumptive Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 30
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 30
4-1
4.0 INTRODUCTION
Implementation of an effective BMP monitoring program is not a straight-forward task. BMPs
by definition are devices, practices, or methods used to manage stormwater runoff. This
umbrella term lumps widely varying techniques into a single category. Some BMPs, especially
nonstructural BMPs are pollution prevention techniques, such as low-impact development
techniques (LIDs), or source control techniques, such as street sweeping. These BMPs cannot be
directly associated with an influent and effluent that isolates their effects. Other BMPs are
strictly governed by diffuse or nonpoint source concepts, such as buffer strips, while larger
regional BMPs such as downstream detention ponds, behave more like intermittent point sources
than traditional diffuse or nonpoint sources. Thus, a wide variety of underlying conditions may
exist, making a one-size-fits-all approach to BMP monitoring infeasible.
Great variability in stormwater properties and the associated runoff complicates BMP monitoring
further. Precipitation varies in time and space. Stormwater pollutants can be carried into the
receiving water system by the precipitation itself (wet deposition) and/or picked up as it flows
across surfaces with or without conveyance by man-made or natural drainage channels. Air
quality, land use, drainage systems, and geology characteristics are all nonuniform, again leading
to temporal and spacial variation concerning stormwater pollutant loads. An effective BMP
monitoring program must incorporate this variability to produce reliable data. Given these
difficulties, it is no surprise that a report produced by the Governmental Advisory Committee in
2000 on current monitoring practices in the U.S. noted that data gaps are prevalent and
“particularly serious for nonpoint sources” (U.S. EPA, 2002a). Nonpoint sources are recognized
as the major contributor to the pollution of the nation’s waters, particularly since regulatory
measures made vast reductions in point source pollution. BMPs are the primary tools used to
mitigate the deleterious effects of nonpoint sources on receiving waters, yet there is little
evidence that BMPs are meeting their projected goals (Strecker and Urbonas, 2001). Therefore,
high quality BMP monitoring programs are an important piece in completing the current picture
of the nation’s water quality and making steps towards improvements.
There are many types of BMP monitoring: implementation, trend, effectiveness, pre-BMP
stormwater monitoring to establish BMP design, project validation, whole system effectiveness,
and compliance monitoring. The goals and objectives of a given monitoring program will dictate
which type of monitoring is appropriate. The focus of this section is effectiveness monitoring.
However, pre-BMP stormwater monitoring to establish BMP design may be the most practiced
type of BMP monitoring, therefore, important issues specific to this type of monitoring will be
identified where applicable. BMP implementation and trend monitoring are covered in detail in
the U.S. EPA report Techniques for Tracking, Evaluating, and Reporting the Implementation of
Nonpoint Source Control Measures - Urban (U.S. EPA, 2001). Although the different types of
monitoring have their own distinct set of characteristics, there is a great deal of overlap between
them. Accordingly, many issues discussed in this section are not exclusive to effectiveness
monitoring.
4-2
4.1 CURRENT MONITORING PRACTICES
There are four main monitoring approaches employed to assess BMP effectiveness (Strecker et
al., 2000). The most popular approach is input/output sampling that is used with new, existing,
or retrofitted structural BMPs. A second BMP monitoring approach is before/after sampling.
This approach can be used in new or retrofit BMP situations, but is most often used with
nonstructural or other BMPs that lack an inflow/outflow. Upstream/downstream monitoring can
be used to assess the impact of a single BMP’s effluent or an untreated stormwater input on its
receiving stream (FHWA, 2000). The final BMP monitoring approach is control watershed
comparison. Although sometimes useful for evaluating nonstructural BMPs where before data
was not collected or structural BMPs without defined inlets (e.g., vegetative filter strips), the
control watershed comparison approach is rarely used due to the difficulty in finding a watershed
with similar contributing factors to serve as the control.
Thorough water and pollution loading budgets are important to a robust BMP monitoring
program. Stormwater flow measurements are of critical importance, yet they are often the cause
of error in efficiency calculations (U.S. EPA, 1999; FHWA, 2001). Dry-weather flows,
groundwater, and direct precipitation can also contribute to both hydraulic and pollutant loading,
although they are rarely considered (Reinelt and Horner, 1995; Rushton, 1998; GeoSyntec and
ASCE, 2002). Long-term studies that include seasonal effects on flow and pollutant patterns are
rare as well (Reinelt and Horner, 1995), even though rainfall/flow and pollutant build-up vary
over time. Other problems often associated with poor quality flow data include a lack of
equipment maintenance and calibration, and the neglect of bypass flows (Clary et al., 2001;
GeoSyntec and ASCE, 2002). These factors can lead to either an over or underestimation of
actual BMP efficiencies.
Collecting representative samples is another important component of a robust BMP monitoring
program. There are many techniques for stormwater sample collection. The simplest is
obtaining a grab sample. Composite samples include time-proportional sampling and three types
of flow-proportional sampling: (1) constant time - volume proportional to flow; (2) constant time
- volume proportional to flow-volume increment; and (3) constant volume - time proportional to
flow-volume increment (GeoSyntec and ASCE, 2002). BMPs that do not have a clearly defined
inflow and outflow, such as biofiltration, greenroofs, and vegetated filter strips, present unique
difficulties in collecting representative samples. Flows into and out of these systems must be
concentrated into a single or otherwise representative outfall before collection (U.S. EPA, 1999).
In addition to being representative in how it was collected, the sample must also be
representative in what parameters are measured. Most BMP monitoring studies sample for
selected chemical and/or physical species such as total phosphorus and total suspended solids
(TSS). However, these parameters do not always provide a representative picture of BMP
effectiveness. Although TSS has proven useful in wastewater monitoring, it has been found
“fundamentally unreliable” for evaluating natural waters (Rushton, 2002). Biochemical oxygen
demand is another traditional wastewater parameter often measured in BMP monitoring
4-3
programs that has been criticized as unrepresentative for stormwater (Rushton, 2002). A key
parameter for evaluating BMP effectiveness that is usually neglected is gross or coarse pollutants
such as floatables or litter debris (e.g., beverage containers, Styrofoam) and organic debris (e.g.,
leaves, grass clippings, and twigs) (England and Rushton, 2003). Biological and physical
indicators of the ecological health of the receiving water may provide more realistic information
on BMP effectiveness, depending on the objectives of the BMP monitoring program, yet these
parameters are not historically measured (U.S. EPA, 1999).
Once data is collected during a BMP monitoring program, it must be analyzed. GeoSyntec and
ASCE (2002) list ten different methods that have been applied to BMP monitoring data to assess
effectiveness. Every method employs its own set of assumptions and as a result produces its own
level of efficiency. This alone is a significant contributor to the wide range of reported BMP
effectiveness. For example, an exercise that applied three of the most common data analysis
methods to the same set of data resulted in percent removals from 48 to 66% (Strecker et al.,
2000). The most commonly used method of data analysis is event mean concentration (EMC)
(U.S. EPA, 1999). An EMC is a statistical parameter representing the average concentration of a
constituent over the course of a single storm event. It can be calculated by dividing the
constituent’s flow-proportional average mass by the total runoff volume. Alternatively, samples
can be collected on a flow-proportional basis to build a flow-weighted composite sample for
analysis. Concentrations of such a composite sample are directly assumed to be EMCs. BMP
efficiency is often then determined based on percent removals between the influent and effluent
EMCs on a storm-by-storm basis. This storm-by-storm pairing of influent and effluent may be
problematic for BMPs with a permanent volume, such as retention (wet) ponds because the
simultaneous inflow and outflow may not be related to the same storm event (Clary et al., 2001).
It is also problematic as it weighs all storms equally. One final reason this analysis may be
inappropriate is that percent removals do not tell the whole picture and therefore can be
misleading. For instance, one study of a detention pond in Florida found that the removal
efficiency actually decreased after a modification to increase the residence time, and thus one
could conclude that the modification resulted in worse performance (Strecker et al., 2000).
However, a closer look at the data revealed that the effluent EMCs after the modification were
less than before the modification, but for unknown reasons the influent EMCs were also lower.
Thus, the comparative differences between the influent and effluent were greater before the
modification than after, leading to the erroneous conclusion. As shown by this example,
emphasis on percent removal is unjustified and it cannot stand alone as a sole measure of BMP
effectiveness (GeoSyntec and ASCE, 2002).
Typically, data analysis of BMP monitoring programs also fail to consider statistical validation
(Strecker et al., 2000). This point has been regarded by some as being “the most frequently
overlooked factor” in a BMP monitoring program (GeoSyntec and ASCE, 2002). Obtaining
statistically valid results begins with good planning to assure that the number of storms sampled
is sufficient to draw valid conclusions at the specified level of confidence. As with any field
experiments, unusual activities within the watershed or uncharacteristic rainfall events cannot be
controlled and if there are too few sampling events, these incongruities can bias effectiveness
4-4
results (Taylor and Wong, 2002b). Thus, the number of sampling events necessary to obtain
statistical confidence may become cost prohibitive. After data analysis is completed, the
statistical significance of a BMP monitoring program’s conclusions should always be reported.
A final downfall of BMP monitoring programs is that details that contribute to the observed
effectiveness of a BMP are often overlooked (Strecker et al., 2000). Watershed characteristics
that may affect BMP performance include watershed area, percent imperviousness, land-use
breakdown, soil types, and rainfall characteristics. BMP design characteristics and maintenance
activities also effect BMP performance. However, it is wise not to rely solely on preconstruction/implementation design plans, as “BMPs will not necessarily be built and/or
maintained as designed and approved” (Taylor and Wong, 2002b). All of these variations result
in a nearly infinite number of combinations that further complicate the evaluation and
comparison of BMP monitoring results (U.S. EPA, 1999).
As the above discussion illustrates, most BMP monitoring studies reported lack the rigor required
to compare and pool individual studies. Pooling of individual studies is necessary to make any
definitive assessments of BMP effectiveness. Rigorous BMP monitoring programs become
complex quickly. Consequently many BMP monitoring programs produce insufficient or
unsound data, in part due to poor experimental design (Marsalek and Kok, 2000). Thus,
extrapolation of BMP effectiveness results to unmonitored BMPs and areas under consideration
for BMP construction/implementation is impossible or unreliable at best. In these situations it is
always best to perform baseline monitoring of the stormwater runoff characteristics and tailor the
BMP design to the specifics of each individual location.
In the past, the field of BMP effectiveness monitoring has suffered from a lack of extensive
experience and good quality guidance. More recently, GeoSyntec and ASCE (2002) produced a
reasonably comprehensive guidance manual for the monitoring of urban stormwater BMP
performance. The main goal of this manual was to assist researchers in producing robust and
consistent BMP monitoring results to expand the National Stormwater BMP Database
(accessible online at http://www.bmpdatabase.org/). In addition to suggesting preferred methods
of sampling, measuring flows, and analyzing data, the manual includes many data collection
tables covering everything from watershed characteristics to significant BMP design parameters
individualized for the most common BMPs.
4.2 PARAMETER SELECTION FOR STRUCTURAL BMP MONITORING PROGRAMS
There is a wide variety of parameters that could be included in a BMP monitoring program.
When developing a monitoring program, the number of possible parameters can be
overwhelming. In this section, potential parameters are introduced under five major parameter
categories. Once familiar with the parameter options, a list of key considerations used to narrow
in on an appropriate set of monitoring parameters is presented.
4-5
4.2.1 Major Parameter Categories
Five major categories of BMP monitoring parameters have been identified: (1) chemical
parameters; (2) physical parameters; (3) biological parameters; (4) hydrological parameters; and
(5) additional contributing factors. Traditional BMP monitoring programs have focused mainly
on water quality and physical parameters (U.S. EPA, 2002a). However, a robust monitoring
program will incorporate some measures from most or all five of the major categories.
4.2.1.1 Chemical Parameters
In the early 1980s, the U.S. EPA’s NURP established a set of nine water quality parameters (and
one physical parameter, TSS, discussed in Section 4.2.1.2) thought to fully characterize urban
stormwater (GeoSyntec and ASCE, 2002). These parameters, listed below, are still the most
widely applied parameters in BMP monitoring programs.
Biochemical Oxygen Demand (BOD)
Total Kjeldahl Nitrogen (TKN)
Chemical Oxygen Demand (COD)
Nitrate + Nitrite Nitrogen (NO2- + NO3-)
Copper (Cu)
Total Phosphorus (TP)
Lead (Pb)
Soluble (or ortho-) Phosphorus (SP)
Zinc (Zn)
Since the 1980s, the BMP scientific community has re-evaluated this list. In 1994, it was
suggested by Strecker that ortho-phosphate was a more useful parameter than soluble phosphorus
because ortho-phosphate approximated the bioavailable fraction phosphorus (Strecker, 1994).
As discussed above, BOD has been criticized as being an inappropriate measure for stormwater.
Individual nitrogen species recycle rapidly, thus their measurement may not be beneficial
(Strecker, 1994; Rushton, 2002). Measurement of total metals may not provide a representative
indication of toxicity because some metals are only toxic in their soluble or unbound states.
Also, hardness is often used in metals standards for toxicity, and without this measurement the
true consequence of metals concentrations cannot be fully assessed (Rushton, 2002). After
reviewing completed BMP monitoring studies for cost-effectiveness and the generation of
meaningful results, Strecker updated NURP’s list of recommended water quality parameters
(Strecker et al., 2000). The updated list reflected the issues discussed above by excluding BOD
and TKN, and adding total organic carbon (TOC), total hardness, ammonia-nitrogen (NH3-N),
and Cadmium (Cd). Both the total and soluble forms of each metal (Cd, Cu, Pb, and Zn) were
recommended for analysis. Additional chemical parameters that are measured less often but can
be relevant in certain situations are oil and grease, polycyclic aromatic hydrocarbons (PAHs),
potentially toxic pesticides and herbicides, and other organic chemicals (Strecker, 1994; Pitt et
al., 1995; NJDA et al., 2000; GeoSyntec and ASCE, 2002; Rushton, 2002).
In baseline monitoring, analyzing for a wide variety of chemical parameters from nutrients to
organics to metals will help identify target parameters for BMP design. It is also important,
particularly for baseline monitoring, to evaluate particulate-associated versus dissolved fractions
4-6
of chemical pollutants, as their modes of removal and transformation within a BMP are often
significantly different.
Monitoring chemical parameters provides specific information about BMP water quality.
Potential effects on receiving waters can be predicted by chemical concentrations and their
quantitative nature provides for a clear method of evaluating BMP performance when inflow and
outflow are compared. However, direct measurement of pollutants is time-specific. Chemical
data is highly variable and thus “100% valid only at the precise moment the sample was taken”
(NJDA et al., 2000). Water quality trends and cumulative effects of chronic, low concentrations
are not readily observed when monitoring chemical parameters alone.
4.2.1.2 Physical Parameters
The physical characteristics of runoff are key to BMP monitoring programs. Not only are the
physical parameters themselves, especially TSS, often the main specified management goal of
BMPs, they also act as an indicator for, and can become carriers of, many of the chemical
pollutants discussed in Section 4.2.1.1 (Ruby and Kayhanian, 2003). As such, TSS is often the
only parameter measured in BMP monitoring programs because it is believed to indicate overall
water quality (Rushton, 2002). However, caution must be used when making this assumption
because the correlation between TSS and other parameters is most often not strong enough to
eliminate the need to specifically address other significant parameters (Strecker, 1994).
Furthermore, TSS may not be a representative measure of the physical nature of runoff because
larger sediments settle out quickly and thus the procedure for TSS measurement will not include
this portion of the sediment mass. In a study supported by the U.S. Geological Survey (USGS), it
was found that TSS measurements may underestimate the actual sediment load by an order of
magnitude or more, particularly in the influent (GeoSyntec and ASCE, 2002). Thus, a new
measurement of sediments, the suspended-sediment concentration (SSC) was developed and is
now part of the USGS policy on water monitoring. SSC differs from TSS solely by the
specification that SSC must use the entire sample, while TSS measurement may (and in most
situations does) use a sub-sample of the originally collected sample. Although SSC is considered
by most to be a more appropriate measure for BMP monitoring programs, TSS measurement
generally is recommended due to its historical use to enable comparison with existing data and
development of trends (GeoSyntec and ASCE, 2002). The freedom to use sub-samples in TSS
measurement also lends itself more easily to the analysis of composite samples often used in
stormwater monitoring. Further, using SSC to determine BMP performance (by percent
removal) with respect to solids can bias the results, while TSS will provide a much more
conservative measure of BMP performance. This bias is less important if other measures, like
average effluent concentrations, are used to determine BMP effectiveness.
Gross solids, such as litter, trash, and other debris, are another significant physical parameter in
stormwater. Although seldom measured, control of gross pollutants is principal to maintaining
the aesthetic value of receiving waters. Since aesthetic quality is the one effect of untreated
stormwater that is easily identifiable and understandable by the public, the effectiveness of BMPs
4-7
in controlling gross pollutants should be included in a robust BMP monitoring plan. In addition
to their direct impact on aesthetics, gross solids may also “degrade aquatic habitat, ...smother
productive sediments, leach harmful pollutants, and cause unpleasant odors” (England and
Rushton, 2003). Measurement of gross solids is difficult because they cannot be collected by the
automatic samplers commonly used in stormwater monitoring (Rushton, 2002). England and
Rushton (2003) have developed a method of monitoring gross pollutants as part of a BMP
maintenance program, measuring volume and dry mass of gross solids during BMP clean out.
Other physical parameters that may be important to monitor include turbidity, particle size
distribution, settling velocity distribution, and accumulated sediments. Particle size distribution
data can be useful to estimating the BMP capture efficiency of suspended solids, particularly if
particle specific gravity is measured (GeoSyntec and ASCE, 2002). Measurement of settling
velocity directly is also an option which is recommended (Field and O'Connor, 1996; GeoSyntec
and ASCE, 2002). Particle size distribution and specific gravity or settling velocity are important
for baseline monitoring to establish BMP design and for influent samples where settleability may
be a significant factor in performance variations between BMPs. Monitoring accumulated
sediments may be important in BMPs that encourage particle settling. Typically, the sediments
represent a sink for both physical pollutants (sediments) and their associated water quality
pollutants. However, under some conditions, accumulated sediments can become a pollutant
source (Kaiserli et al., 2002). In areas with high heavy metals loads, the accumulated sediments
may actually become toxic and subject to hazardous waste regulations (WEF and ASCE, 1998).
Accumulated sediments can be measured by coring, grab sampling, or using sedimentation traps
that have the added benefit of allowing estimates of sedimentation rates (Pettersson, 2001).
One final item that can be characterized as a physical parameter is physical changes to the
receiving water channel caused by stormwater (WEF and ASCE, 1998). Such changes include
stream down-cutting and are primarily the result of changes in the stream hydrology that may be
altered by the BMP. U.S. EPA’s Rapid Bioassessment Protocols include a physical habitat
assessment component that may be used to monitor physical changes to the receiving water
channel over time (U.S. EPA, 1999).
4.2.1.3 Biological Parameters
Biological parameters can be divided into two main groups, those that BMPs strive to prevent
(i.e., direct measurement of organisms associated with pollution), and those that BMPs strive to
protect (ecological effects on receiving waters). The former is still relatively rare in BMP
monitoring programs (Rushton, 2002). Pathogenic microorganisms present in stormwater can
lead to fish kills in extreme cases and more often beach closings. The impact of BMPs on
pathogenic microorganism such as bacteria, viruses, Cryptosporidium and Giardia, are not well
characterized at present due to limited monitoring data. Indicator organisms such as fecal
coliform, which are traditionally used in the wastewater industry, may not be representative for
stormwater applications (O'Shea and Field, 1992; Rushton, 2002). However, total and fecal
coliform remain the basis of most permits and state water quality standards, and are the reasoning
4-8
behind most beach closings. Thus, measurement of these indicator organisms provides practical
information.
The overall goal of any BMP implementation is ultimately to preserve or restore the health of the
nations’s waterways. Therefore, monitoring ecological effects downstream of a BMP is
considered by some a better indicator of BMP effectiveness than water quality parameters alone
(Clary et al., 2001). This is especially true when monitoring for a long-term response since
ecological effects generally reflect trends in water quality, barring acute effects of obvious
catastrophic events (NJDA et al., 2000). Ecological effects on receiving waters can be assessed
in two ways, toxicity testing of the BMP effluent and in-stream indices. Both provide
information for assessing the combined impact of multiple stressors (U.S. EPA, 1996).
Whole water toxicity testing for BMPs removes a sample of a BMP effluent stream, places many
organisms of a single species directly into the sample and monitors the mortality rates, growth
rates, and other changes in behavior or overall health. The intermittent nature and variability in
flow rates and concentrations unique to wet-weather flows requires special protocols for a true
assessment of toxicity potential. Herricks (1996) has developed toxicity protocols for wetweather flows that take into consideration the timescale issues associated with pulsed inputs of
stormwater runoff. The Microtox® toxicity-screening procedure has also been used to evaluate
stormwater (Pitt et al., 1995). Toxicity testing has the advantage over in-stream indices of being
performed in a controlled laboratory environment using indicator organisms whose growth
characteristics are well known, such as Daphnia pulex or Pimephales promelas (NJDA et al.,
2000). The test usually will show negative effects, if present, within 24 to 72 hours. When
combined with water quality parameter concentrations, toxicity testing can provide an indication
of pollutant bioavailablity (Marsalek and Kok, 2000). However, the tests can be costly and their
results may be highly variable (Marsalek and Kok, 2000; NJDA et al., 2000).
In-stream indices can be developed from analysis of fish, benthic macroinvertebrates, and plant
communities (U.S. EPA, 2002a). Fish are useful long-term indicators because their longer life
span (3 to 4 yrs) allow for bioaccumulation of toxins in the organism’s tissue (NJDA et al.,
2000). Fish communities can be monitored for individual species abundance, indication of
deformities or illness, such as discoloration, and pathologic parameters, like presence of tumors
or bioaccumulated toxin concentrations. Fish of the highest trophic level are particularly useful
because “they tend to integrate changes in lower trophic levels, ...[thus reflecting] overall
ecosystem condition” (U.S. EPA, 2002a). Including fish in a BMP monitoring program also can
be a good way to muster public interest and support. Unlike other biological indices, fish are an
organism the public can easily identify as beneficial to the watershed, especially in areas where
sportfishing is popular. However, in some cases the receiving water may be too small to support
a fish community. Benthic macroinvertebrates, such as snails and insect larvae, have limited
mobility and shorter life spans than fish. Thus, they provide better site specific information
which may enable source location and their reaction times are faster (NJDA et al., 2000).
Identification of macroinvertebrates is more difficult and time consuming. In addition there are
seasonal trends in species types that can vary from year to year. These issues limit its
4-9
applicability. Methods for assessing fish and benthic communities, detailed in U.S. EPA’s Rapid
Bioassessment Protocols, may be suitable for assessing BMP effectiveness (U.S. EPA, 1999).
Finally, plants, being at the bottom of the food chain, can provide a direct link to pollutant levels.
An example of a plant indicator is the presence/amount of algae attached to rocks (U.S. EPA,
2002a).
All three in-stream indicators (fish, benthic macroinvertebrates, and plant communities) share
several disadvantages. First, in-stream indices are not specific to the effects of BMPs. Many
other urban stressors contribute to a stream’s biological condition, including “disruption of
physical habitat, alteration of hydrologic patterns, introduction of non-indigenous biota, and
widespread alteration of the landscape” (U.S. EPA, 2002a). Depending on the size of the stream
as compared to the volume of BMP effluent, simple dilution effects may obscure any BMP
effects. Additional factors, such as fish harvesting/stocking, rainfall or drought patterns,
interfering upstream conditions, or possible groundwater interferences, can overshadow the
effects of any single BMP (Strecker and Urbonas, 2001). Second, the lag between introduction
(or removal) of the stressor and observed biological response is long enough that in-stream
indices are inappropriate for evaluation on a storm-by-storm basis (U.S. EPA, 1996). Third, instream parameters exhibit a large degree of seasonality. Lastly, results of in-stream biological
parameters vary greatly depending upon the characteristics of the sampling location. Streams
may be both horizontally and vertically stratified due to primary and secondary currents. It is
particularly difficult to obtain a representative sample of chemically reactive or particulate
associated pollutants whose concentration will vary with distance from the BMP outfall. The
pollutant’s reaction rates (chemical and settling) in relation to the stream velocity must be taken
into account when choosing the optimum sampling location for reactive pollutants. Other
localized factors that may affect in-stream pollutant concentrations include riffle vs. pool and the
percentage of area receiving full sunlight (U.S. EPA, 1996). Resuspension of pollutants
associated with previously settled sediments is another confounding factor. For these reasons,
sampling location characteristics should be well documented.
4.2.1.4 Hydrologic and Hydraulic Parameters
The foundation of a good BMP monitoring program, both baseline and effectiveness, is an
accurate and representative measurement of precipitation and stormwater flow data (U.S. EPA,
1999). The hydraulic alteration of stormwater associated with human activity such as increases
in peak runoff rates and total runoff volumes is a problem in and of itself, causing downstream
flooding, lower base flows, increased erosion, and habitat destruction (NJDA et al., 2000). The
role of BMPs in mitigating these effects are a substantial and historic component of their
intended function.
Defining storm and runoff hydrology is not a trivial matter. A number of hydrologic and
hydraulic parameters such as antecedent conditions, pattern of precipitation intensities,
precipitation durations, and total precipitation volumes, runoff rates, durations, and total volumes
into and out of the BMP, all contribute to “make each storm a unique event”(Church et al.,
4 - 10
1999). Additional parameters such as possible bypass or overflow volumes and influences of soil
infiltration, groundwater, dry-weather flows, and evaporation add another layer of complexity.
All this variability translates into a great deal of inconsistency in reported BMP effectiveness.
Therefore, measurement of hydrologic and hydraulic parameters is key to interpreting BMP
monitoring data and predicting effectiveness.
The necessity of some hydrologic and hydraulic parameters is dependent on the BMP type. For
BMPs designed to encourage infiltration such as infiltration basins, flow reductions are the
primary mechanism of stressor alleviation. A reasonably accurate water balance including all
inflows and outflows and antecedent conditions and precipitation intensities, both of which will
affect infiltration capacity, are essential for these types of BMPs. Accurate flow measurements
are especially important since they are often used as a basis for collecting water quality samples
and required for the calculation of pollutant loadings. Precipitation volumes and durations are
often significant in defining storm events appropriate for monitoring, although it is optimum to
monitor all wet-weather events continuously.
The USGS report, Basic Requirements for Collecting, Documenting, and Reporting Precipitation
and Stormwater-Flow Measurements prepared by Church et al. (1999), is an excellent source of
guidance for collecting this type of data. Information on the actual collection of hydrologic and
hydraulic parameters is discussed further in Section 4.5.2.3.
4.2.1.5 Contributing Factors
In addition to chemical, physical, biological, and hydrologic/hydraulic parameters, many other
contributing factors determine the overall performance of a BMP. Watershed characteristics
such as size, geographical location, land use, percent imperviousness, topography, soil
characteristics, and upstream nonstructural BMPs should be collected during a comprehensive
BMP monitoring program (U.S. EPA, 1996; GeoSyntec and ASCE, 2002). BMP design
characteristics are also important contributing factors. Variables such as length, width, depth,
storage volume, inlet and outlet design, vegetation, and bottom lining or soil infiltration
characteristics should be documented. The complete set of BMP design characteristics is
specific to the type of BMP being monitored. GeoSyntec and ASCE (2002), in their Urban
Stormwater BMP Performance Monitoring: A guidance manual for meeting the national
stormwater BMP database requirements have included detailed tables and photocopy-ready
forms of BMP characteristics useful for evaluating effectiveness for many of the more common
BMPs. Additional BMP characteristics include age and maintenance practices.
Contributing environmental characteristics, such as temperature, dissolved oxygen (DO), solar
intensity, and pH, are also useful in the analysis of monitoring data. Temperature can affect
process reaction rates within the BMP. Also, heat is specifically mentioned as a pollutant in the
Clean Water Act’s definition of pollutant (Parikh, 2003). Thus, increased temperatures,
produced by areas collecting runoff from large paved areas or by the quiescent water of retentiontype BMPs during dry periods, could be a pollutant (NJDA et al., 2000). DO concentrations may
4 - 11
be an important parameter to monitor in detention-type systems. Pollutants such as phosphorus
and some metals (Cd, Cu, Pb, and Zn) that settle to the bottom with suspended solids may leach
into the water column during periods of low DO concentrations (Rushton, 1998). As with
temperature, low DO can also be considered a pollutant. The oxygen dependant organisms
within a receiving water, such as fish, can be severely impaired or killed by prolonged low DO
conditions. Hardness, as discussed earlier in Section 4.2.1.1, is another contributing
environmental characteristic that is particularly useful for metals analysis.
4.2.2 Key Considerations for Selecting Appropriate Parameters
The number of possible parameters that may be measured in a BMP monitoring program is
extensive. It is often impractical to measure all the parameters discussed in Section 4.2.1, and
there is no “one-size-fits-all” set of parameters that will satisfy the objectives of every monitoring
program. The planning phase of a BMP monitoring program must include the selection of
appropriate parameters. The following questions are a list of key considerations that may be
useful during the parameter selection process.
What parameters are required to meet the monitoring program objectives and goals?
If the monitoring program objectives are well defined, this may be the only question that needs
asking. The objectives and goals will depend in part on the given type of monitoring program.
Parameters that are appropriate to meet the objectives and goals of a baseline monitoring
program to establish BMP design may be different from those of an effectiveness monitoring
program. Further, BMPs may be implemented for many reasons, which could include regulatory
compliance, total maximum daily load (TMDL) compliance, protection of sensitive ecosystems,
etc. These reasons typically define the monitoring objectives and goals and, in turn, the list of
appropriate parameters (Caltrans, 2000). For example, if a BMP is constructed to aid in the
compliance with a TMDL for phosphorus, phosphorus must be measured.
What resources are available for completing monitoring objectives?
Resources include money, personnel, and time. There is often a balance between the number of
parameters monitored and the number of events for which they can be measured that is usually
driven largely by available resources. In most cases, limiting the number of parameters to those
most significant to the monitoring objectives, will allow more storm events to be sampled,
leading to better supported conclusions (GeoSyntec and ASCE, 2002).
Do any regulatory or legal requirements apply to the BMP or its receiving waters?
Those parameters specified in any regulatory requirements or court-ordered legal requirements
must be included in the BMP monitoring program. BMP specific regulatory considerations
include the use of the 80% reduction in TSS rule and the National Pollutant Discharge
Elimination System (NPDES) Stormwater Phase II Regulations. The coastal zone management
measures guidance require an annual average reduction in TSS loadings of 80% in construction
areas (NJDA et al., 2000). The U.S. EPA’s Phase II requirements include provisions to track the
implementation of five minimum stormwater management measures: public education and
4 - 12
outreach, public involvement, illicit discharge detection and elimination, reduction of
construction site and post-construction runoff, and pollution prevention/good housekeeping
activities. The U.S. EPA included no specific requirements for chemical (or biological)
monitoring information. Thus, some municipalities have developed a BMP monitoring program
that compiles qualitative data only about both structural and nonstructural BMPs to satisfy their
permit requirements (Hillegass, 2003). A monitoring program of this type primarily tracks
contributing factors, e.g., the type of BMP used and the size of its catchment area.
Applicable surface water quality standards of the receiving water should be reviewed before the
final parameter selection. For example, if the water quality criteria specify levels for total metals
and the monitoring plan only calls for soluble metals or vice versa, the data may not be able to
answer key questions concerning the effectiveness of the BMP.
Are existing monitoring data available?
Data from previous monitoring programs or screening studies can be useful in identifying
appropriate parameters for a successful BMP monitoring program (as well as establishing
estimates of the number of samples need to detect desired levels of change). Constituents
prevalent in existing monitoring data, especially those above levels of concern and those targeted
by the BMP should be included, whereas those constituents rarely detected generally can be
eliminated unless other circumstances exist, such as a recent change in the catchment area, or a
regulatory or legal requirement (Caltrans, 2000). Selecting appropriate parameters based on
existing data usually leads to a cost-effective BMP monitoring program. Screening studies to
ascertain what pollutants might be of concern/interest can be valuable during the design phase of
a BMP monitoring program (GeoSyntec and ASCE, 2002).
What are the prevailing land uses in the catchment area?
Different land uses have been associated with different types of pollutants and potential pollutant
sources. Knowledge of the prevailing land uses in the target catchment area can help identify
parameters likely to be in the wet and dry-weather runoff. For instance, if the catchment area
includes a large amount of highway or industrial yard runoff, parameters such as metals, PAHs,
and industry specific toxic substances are likely to be of concern (Caltrans, 2000). However, if
the catchment area is heavily influenced by construction or agricultural activities, sediment
pollution is typical. In catchment areas subjected to combined or sanitary sewer overflows,
pathogens may be of importance.
What are the beneficial uses and impairments (if any) of the receiving water?
Beneficial uses or impairments of receiving waters are often the underlying reason behind BMP
implementation. Monitoring programs can be used as verification that the BMP is fulfilling its
intended purpose. It has been recognized that “in many instances the water quality problem will
directly indicate what variables should be monitored” (U.S. EPA, 1996). For instance, if the
BMP discharges near a public beach, pathogens or bacterial indicator monitoring will be
important. Or, if the BMP discharges to a stream that supports a healthy game fish population,
then in-stream biological indicators may be useful.
4 - 13
Are there any parameters that are particularly useful for evaluating the type of BMP being
monitored?
Some parameters will be more important than others, depending on the type of BMP being
monitored. For example, soil permeability may be an important parameter for infiltration BMPs,
but less important for ponds, especially those that are concrete or clay lined. Alternatively,
settleable solids may be an important parameter for wet ponds, but less important for infiltration
BMPs. Another example is particle settling velocity distribution for BMPs that depend on
removal by inertial separation of suspended solids, and particle size distribution for BMPs that
depend on filtration (Field and O'Connor, 1996).
Are there any contributing factors that would be useful in interpreting data from the primary
parameters selected?
Collection of supporting data can greatly enhance the usefulness of BMP monitoring data
(Caltrans, 2000). The most striking example is the measurement of hardness when metals have
been selected as a primary parameter, as discussed above in Section 4.2.1.1. Other examples of
contributing factors are given in Section 4.2.1.5.
Are the parameters typically monitored constituents?
Whenever possible, the parameters selected should be typical (e.g., those typically found in urban
stormwater runoff at levels of concern). Typically measured parameters are most likely to have
straightforward and reliable methods of analysis (GeoSyntec and ASCE, 2002). Their data is
easily comprehensible by those in the environmental profession, which gives them a broader
usefulness. If typical parameters are measured, the data will be easily comparable to other
studies with similar circumstances and will tend to be more economical overall.
4.3 MONITORING NONSTRUCTURAL BMPS
According to an international survey, the use of nonstructural BMPs is on the rise and the trend is
expected to continue (Taylor and Wong, 2002a). Yet, there is a paucity of high-quality studies
on their effectiveness in improving the quantity and quality of stormwater. This lack of
monitoring of nonstructural BMPs has been cited on more than one occasion “as a major
impediment to their adoption” (Taylor and Wong, 2002a). Monitoring of nonstructural BMPs is
inherently difficult for many reasons. The most significant hindrance to monitoring nonstructural
BMPs is that many of them rely on behavioral change. Behavioral changes are difficult to
measure for the following reasons: (1) direct observation of behavior is usually not an option for
privacy reasons; (2) self-reported behaviors are often vastly different from the truth; (3) people’s
attitudes toward environmental issues do not often translate into actual behaviors; (4) developing
suitable quality controls is difficult; and (5) there is a time lag between awareness about an issue
and actual behavioral change, which could require up to ten years (Taylor and Wong, 2002a;
Taylor and Wong, 2002b).
4 - 14
A second reason for the difficulty monitoring nonstructural BMPs is the lack of defined inflows
and outflows. Therefore, the effects of nonstructural BMPs cannot be isolated for direct
measurement. Taylor and Wong (2002a) conceded that monitoring the effects of some
nonstructural BMPs (e.g., public education and outreach) on stormwater quality is “virtually
impossible...or at best can be evaluated using [long-term] trend analysis.” A third difficulty is
nonstructural BMP programs are often backed by substantial resources and thus, strong pressures
to report positive results may exist. This has the potential to skew the results of any monitoring
that has been conducted (Taylor and Wong, 2002b). One final word of caution regarding the
monitoring of nonstructural BMPs is that the effectiveness results collected may not be
transferable to other communities (Taylor and Wong, 2002a). People’s behaviors are shaped by
social, educational, economical, and regional factors; therefore, what works in one place, may not
be universally effective.
Despite its difficulties, nonstructural BMP monitoring is feasible using a number of approaches.
Taylor and Wong (2002a) identified seven approaches to evaluating nonstructural BMPs. Many
of the approaches identified are qualitative in nature, such as monitoring public awareness of
target issues and monitoring self-reported behaviors. Obtaining quantitative monitoring data on
nonstructural BMPs can be accomplished in three ways. The first is through modeling. Through
the use of models, qualitative data can be manipulated into quantitative estimates. For example,
the percentage of people who claim to have changed their use of lawn chemicals may be
translated into an estimated reduction of herbicides in runoff. The second way is the direct
measurement of loads prevented from entering the stormwater. This approach has limited
applicability to BMPs such as street sweeping and catchbasin cleaning where pollutants are
collected and can be weighted. The third way to obtain quantitative data for nonstructural BMPs
is long-term trend monitoring of a downstream, end-of-catchment system, such as in-stream
parameters of the receiving water (see watershed monitoring below), a structural BMP influent,
or other end-of pipe system. Although the effects of specific BMPs cannot be isolated, data
collected before and after nonstructural BMPs implementation, or using a reference watershed
can produce valuable quantitative data in support of the effectiveness of nonstructural BMPs.
4.4 WATERSHED MONITORING
A watershed provides a contained unit of receiving waters with ecologically significant
boundaries (WEF and ASCE, 1998). It will cumulate the impacts of human activities such as
urbanization, which has been directly linked to receiving water degradation (U.S. EPA, 1999).
The concept of best management practices and all the related techniques and technologies were
developed with the ultimate goal of minimizing the impacts of urban runoff (both wet and dry)
on receiving waters. Therefore, watershed monitoring may produce a more accurate measure of
the true value of BMPs. For some structural BMPs with no defined inflow and outflow and for
nonstructural BMPs, watershed monitoring will be the only valid monitoring option. Also,
watershed monitoring has the ability to combine the effects of a wide-reaching stormwater
management program. When these programs involve a number of nonstructural and structural
4 - 15
BMPs, it can be more economical to capture the effects of all BMPs in one watershed monitoring
program than to attempt to monitor each BMP individually. However, a disadvantage is that the
effect of any one BMP becomes nearly impossible to ascertain. Although watershed monitoring
typically refers to in-stream monitoring of receiving waters, the same techniques may be applied
to pipe systems or “sewersheds” when appropriate.
Of the four BMP monitoring approaches discussed in Section 4.1, only one, the influent/effluent
approach, truly isolates the BMP’s effects on water quality. The remaining three effectiveness
monitoring approaches, before/after, upstream/downstream, and control (or reference or paired)
watershed, can be applied to the watershed as a whole. An example of the before/after
monitoring approach is the Englesby Brook monitoring project being conducted by the USGS in
Burlington, Vermont (Medalie, 2000). Several BMPs were scheduled to be constructed for
stormwater management. An in-stream monitoring program was initiated two years before
completion of the first BMP in 2002. The monitoring is planned to continue beyond completion
of the final BMP for a total of seven years.
The control watershed approach is the most difficult to apply successfully because of the many
variables involved (GeoSyntec and ASCE, 2002). Typically, control watersheds can differ in
total area as long as the ratio of pervious to impervious areas closely approximate that of the
target watershed, and other watershed characteristics also correspond relatively well. Otherwise,
“as the characteristics of the two watersheds diverge...the noise in the data becomes greater than
the signal” (GeoSyntec and ASCE, 2002).
Watershed monitoring also may be used for trend or compliance monitoring where sample
characteristics would be compared over time or with designated-use goals or water quality
standards (GeoSyntec and ASCE, 2002). A good example of this type of watershed monitoring
is the U.S. EPA’s Environmental Monitoring and Assessment Program (EMAP) (U.S. EPA,
2002a). EMAP uses a probability-based sampling design to obtain a statistically-valid picture of
the ecological condition of statewide water bodies. Currently, planning is underway to apply the
key concepts of EMAP to monitor BMPs on a watershed scale. One disadvantage to using the
EMAP approach for BMP monitoring is the need for a large participating area to ensure a
representative sample, which drives up the cost of implementation.
Special attention needs to be paid to measurement location in watershed monitoring, as there are
no fixed sampling locations as in input/output sampling (FHWA, 2000). Sensors and flow
metering devices must be placed in a location at least twenty channel widths downstream from
any channel bends. The intake for automatic water sampling equipment should be located in a
portion of the waterway that is relatively well-mixed in cross-section. To prevent misleading
results from bedload contamination of the samples, the intake also should be fixed at a height of
100 to 200 mm (4 to 8 in.) above the streambed (FHWA, 2000). Even so, some biases can occur.
Watershed monitoring has its challenges. Because receiving waters are influenced by many other
stressors (e.g., natural weather patterns), the effects of stormwater management programs may be
4 - 16
obscured. The slow biological response to improving conditions and dilution effects that reduce
sensitivity to chemical improvements may make detection of statistically significant changes
impossible (GeoSyntec and ASCE, 2002). This, compounded by annual, seasonal, and even diel
(Brick and Moore, 1996) variability, requires long-term continuous monitoring. A minimum of
two to three years is recommended for watershed monitoring programs (FHWA, 2000). Finally,
from a planning and policy point of view, watershed monitoring may not be a viable option
because watersheds do not usually follow political boundaries, and involvement of all necessary
municipalities may not be possible (WEF and ASCE, 1998).
4.5 DEVELOPING A BMP EFFECTIVENESS MONITORING PROGRAM
Developing a BMP monitoring program that produces useful results takes a great deal of effort
before any samples are taken. The Federal Highway Administration (FHWA) (2000) produced a
well presented guide to the development of a BMP monitoring program. The agency organized a
BMP effectiveness monitoring program into four phases:
(1) planning phase,
(2) design phase,
(3) implementation phase, and
(4) evaluation phase.
This section follows the FHWA’s organization relatively closely, but each section has been
supplemented with information from additional sources. The FHWA’s document is
recommended for anyone involved in developing a BMP monitoring program because each phase
is supported with useful examples to help conceptualize how the guidelines are put into practice.
4.5.1 The Planning Phase
The planning phase is a critical first step in developing an efficient BMP monitoring program. In
the planning phase program goals are defined, background information is collected, and
resources are identified (FHWA, 2000). Using this information, specific project objectives can
be formulated. These objectives form the framework within which the remainder of the BMP
monitoring program is designed, implemented, and evaluated. The U.S. EPA (1996) recognizes
that well defined goals and objectives are “the most fundamental step in the development of a
monitoring plan.”
4.5.1.1 Defining Program Goals
Monitoring goals are broad statements that cover the issues of concern for a particular
stormwater management monitoring program (FHWA, 2000). Examples of goals for BMP
monitoring programs include, evaluation of: (1) flood control; (2) changes in runoff volume and
release rates; (3) pollution reduction by a specific BMP or extended stormwater management
program; (4) the possibilities of stormwater harvesting from the BMP; and/or (5) BMP longevity
or maintenance requirements (Argue, 1995; FHWA, 2000; GeoSyntec and ASCE, 2002). Each
4 - 17
goal often requires a different set of assessment methods and data needs. This section focuses
primarily on concerns related to the goal of evaluating the pollution reduction of a specific
structural BMP.
4.5.1.2 Collecting Background Information
Background information is essential to formulating appropriate objectives in a BMP
effectiveness monitoring plan. Background information includes, site and catchment area
characteristics, BMP design characteristics, receiving water reference conditions, site hydrologic
assessment, and any existing data. Site and catchment characteristics and BMP design
characteristics also are considered critical to evaluate efficiency differences between BMPs
(Strecker et al., 2000). As mentioned earlier in Section 4.2.1.5, thorough tables of these
contributing factors can be found in GeoSyntec and ASCE (2002). Upstream BMPs are an
important component of catchment characteristics that may affect BMP performance for certain
parameters (GeoSyntec and ASCE, 2002). Receiving water reference conditions and existing
data can be used as a point of comparison and for identifying areas of concern (NJDA et al.,
2000). A hydrologic assessment to determine a site’s distribution of storm size and frequency
can help to develop an appropriate temporal scale for the BMP monitoring program (FHWA,
2000). The hydrologic assessment also can determine the possibility of groundwater influences.
4.5.1.3 Identifying Project Resources
Identification of project resources is necessary to assure that the monitoring program can be
attained realistically. Limitations of project resources will ultimately bound the overall scope of
the program objectives. Modeling may be considered to fill in gaps when resources are short,
although good field data is preferred (WEF and ASCE, 1998). Examples of areas where
modeling may be used in lieu of monitoring data are the use of measured BMP effluent data to
estimate receiving water concentrations instead of monitoring the receiving waters themselves
and flow modeling for load determination instead of actual flow monitoring. Two additional
solutions to limited resources suggested by GeoSyntec and ASCE (2002) are to break down the
program goals into smaller and more manageable questions and, if possible, limit the number of
parameters monitored before cutting the number of samples collected.
A more creative way to help keep costs low while gathering an extensive database on BMP
effectiveness is to enlist the help of volunteers. This option has the added benefit of counting
toward compliance with two of the minimum control measures of the NPDES Stormwater Phase
II Regulations, public education and outreach and public participation and involvement (Baxter,
2002). Rhode Island and Connecticut currently boast successful volunteer monitoring programs.
After a short training period, approximately 250 volunteers are given the supplies to perform
weekly or biweekly monitoring at over 120 sites throughout the two states. In Alpharetta, GA, a
similar program relies on volunteers to collect and deliver samples to the lab, but for safety
reasons the volunteers do not actually handle any of the chemicals used for their water quality
monitoring assessments (Baxter, 2002). The most problematic issues with employing volunteers
4 - 18
as part of a monitoring program is quality assurance and continuity. It is imperative that a good
quality assurance plan is in place and that volunteers have training on the information. However,
even if these measures have been taken, adherence to the quality assurance procedures are likely
to be more difficult to enforce when using volunteers. It is also more difficult to keep a
monitoring program operating with regularity when relying on volunteer help.
4.5.1.4 Formulating Monitoring Objectives
Once the first three steps of the planning phase have been completed, specific monitoring
objectives can be set. The FHWA (2000) suggests looking at the program goals as the question
that needs to be answered and program objectives as what needs to be done to answer that
question. Knowledge of background information and available resources should allow the jump
to be made between program goals and program objectives. Clear and specific objectives are
important, as they will drive the program design phase.
4.5.2 The Design Phase
The design phase translates the objectives into an action plan. Issues that need to be defined
include monitoring approach, parameter selection, hydrologic data collection protocols, water
quality data (including chemical, physical, and biological parameters) collection protocols,
identification/selection of equipment and materials, and quality assurance/quality control
(QA/QC) initiatives. The product of the design phase should be a quality assurance project plan
(QAPP) that lays out these details, providing a pathway for meeting the monitoring program
objectives. This phase is the foundation of the project and should be given considerable
attention. A poorly-designed monitoring program could produce misleading data and erroneous
conclusions, resulting in great deal of wasted time and money (GeoSyntec and ASCE, 2002).
The following subsections are brief discussions of key issues included in the design phase.
4.5.2.1 Monitoring Approach
The four main monitoring approaches have already been discussed in Sections 4.1 and 4.4. The
influent/effluent approach to BMP monitoring is used most often. It has the advantage of being a
straightforward approach, isolating the effects of the BMP, usually costing substantially less than
other approaches, and having a short time requirement to reach significant findings (FHWA,
2000). However, receiving water response has been indicated as “a better gauge of long-term
BMP effectiveness...” (Clary et al., 2001). Three so-called watershed approaches monitor the
receiving water response. These approaches, upstream/downstream, before/after, and control (or
reference or paired) watershed, are discussed in length in Section 4.4. Typically, the
influent/effluent approach is the optimum choice, with watershed approaches reserved for BMPs
that do not have a defined input and output, or for cases in which receiving water quality is part
of the monitoring program objectives.
4 - 19
4.5.2.2 Parameter and Methods Selection
Selection of appropriate parameters is an important step. As such, parameter selection has been
expanded into its own section (4.2), and therefore will not be covered here in detail. A few
additional issues regarding methods selection should be addressed. First, the selected methods
should be standard. Using unconventional methods makes comparisons between monitoring
programs unreliable, diminishing the value of the data collected (Ruby and Kayhanian, 2003).
Second, methods that require pre-acidification of sample bottles can be problematic, especially
when using automatic samplers (Rushton, 2002). Experience with these methods showed that
the preservatives were prone to evaporation and sample size estimation was difficult. Third, the
detention limits of the methods chosen should reflect the expected ranges of the pollutants of
concern. Although a data set of mostly non-detects is not completely useless, actual numbers are
always more desirable.
4.5.2.3 Hydrologic and Hydraulic Data Collection
A robust BMP monitoring program begins with hydrologic monitoring. Rainfall intensity,
duration, and total precipitation volume per storm can be monitored with recording precipitation
gauges (Church et al., 1999). Non-recording precipitation gauges are not recommended as they
do not provide the intensity-duration timing necessary to obtain a clear picture of the pollutant
mobilization energy of the storm. Three types of precipitation gauges are generally acceptable,
weighting, float, and tipping-bucket for rainfall (Church et al., 1999). Weighting gauges and
heated tipping-buckets are the acceptable methods for snow, which may be an important issue for
BMP monitoring studies in northern climates. Precipitation gauges are most useful when located
near the water quality monitoring stations, especially when used to initiate autosamplers. Using
precipitation to initiate autosamplers is somewhat problematic because of the spacial variability
inherent in rainfall. Also, depending on land-use and seasonal conditions (e.g., snow event),
precipitation does not always guarantee runoff. For these reasons, an increase in flow rate is
usually a better trigger for autosamplers. Care should be taken so that the gauge is not under
trees or other vegetation that will intercept the rainfall (GeoSyntec and ASCE, 2002). It also may
be advantageous to collect a precipitation sample for water quality monitoring. This becomes
especially important if nitrogen species are of concern because “much of the ammonia and nitrate
in stormwater are deposited directly in rainfall” (Rushton, 1998). For BMPs with a large surface
area, the input of precipitation falling directly on the surface of the BMP may be significant.
Hydraulic data is essential for accurate determination of pollutant loadings (GeoSyntec and
ASCE, 2002). Concentration data alone does not convey the full impact of stormwater or BMP
effluent impact on the receiving water. A complete hydraulic picture includes flow rates and
volumes of BMP inflow and outflow during a storm event, as well as dry-weather flows,
bypasses or overflows, and groundwater flows. Flow rates can be estimated in a number of ways,
but the method recommended by GeoSyntec and ASCE (2002) for BMP monitoring programs, is
through the combination of a primary control device (flume or weir) and a secondary control
device (float gauge, bubbler, pressure transducer, ultrasonic level sensor, ultrasonic uplooking,
4 - 20
radar/microwave sensor, and pressure probe). Primary control devices have channel geometries
that have been calibrated against the depth of the water flowing through (flume) or over (weir)
the device to develop a relationship between water depth and flow. The secondary device
measures the water depth. GeoSyntec and ASCE (2002) contend that the accuracy of primary
control devices makes any additional costs associated with installing the devices worthwhile
because errors in flow measurements will propagate through data analysis (GeoSyntec and
ASCE, 2002). Weirs are generally more accurate and easier to construct and install than flumes
(GeoSyntec and ASCE, 2002). Primary devices do have some difficulties. By constricting flow,
they cause water to backup, allowing for sedimentation (Church et al., 1999). This will change
the physical and chemical characteristics of the samples, leading to inaccurate conclusions.
Flumes are less likely to cause this problem, due to their self-flushing design, and although they
may be more difficult to construct and install, when done correctly, accuracies are on par with or
greater than weirs (Church et al., 1999). Therefore, if a primary device is to be used, Church et
al. (1999) recommends flumes. Guidelines for selecting an appropriate secondary device can be
found in GeoSyntec and ASCE (GeoSyntec and ASCE, 2002).
Regardless of how flow is measured, caution should be taken when comparing results collected
using different methods. Individual flow measurement methods can have considerable variability
and bias (Church et al., 1999). Flow devices should be calibrated frequently to ensure
measurements are as accurate as possible.
Measurement of inflow and outflow data during a storm event is typical, but dry-weather flows
not associated with hydrologic monitoring can be important, as well. It had been recognized that
over the long-term, dry-weather pollutant loads often dominate those resulting from wet-weather
flows for many pollutant types (Pitt et al., 1993). Therefore, flow records should be inspected
for the frequency and volume of dry-weather flows. If dry-weather flows are a significant part of
total flow through the BMP, water quality samples of dry-weather flows should be taken. Flow
through bypasses or overflows must also be considered. Bypasses and overflows can
significantly affect the efficiency of a BMP system as a whole (GeoSyntec and ASCE, 2002).
Disregarding these flows during data analysis will result in misleading conclusions. If direct
flow measurements are not possible, accurate measurements of inflow and outflow data may be
used to estimate bypass or overflow volumes through mass balance calculations. Pollutant
loadings, instead of concentrations, will provide more accurate efficiency evaluations in these
situations.
In addition to surface hydraulics, subsurface hydraulics also may be necessary. Groundwater
contributions are especially important for BMPs located at or near the groundwater table (U.S.
EPA, 1999). Groundwater contributions can be directional and, thus, may require the installation
of both shallow and deep piezometers to estimate horizontal and vertical flow, respectively
(Reinelt and Horner, 1995). If groundwater is found to have significant inflow to the BMP,
water quality samples of the groundwater should be taken. Reinelt and Honer (1995) found that
80% of the total phosphorus loading to a wetland BMP was attributable to groundwater inflow.
Had groundwater sampling not been conducted, efficiency calculations would have erroneously
4 - 21
concluded that the wetland itself was releasing phosphorus when in fact it was reducing
groundwater concentrations. This example indicates the importance of comprehensive flow and
pollutant budgets in BMP monitoring programs.
4.5.2.4 Water Quality Data Collection Protocols
Water quality data in this section refers to all chemical, physical, and biological parameters that
are used to characterize the quality of water. This data is generally the heart of the BMP
monitoring program. It has been suggested that differences in sample collection may be the
largest source of variation in BMP performance and as such thoughtful design of water quality
data collection protocols is imperative (Strecker, 1994). The sampling location, sampling
method, sampling frequency, and sample representativeness all need to be considered in the BMP
monitoring design. Each of these four elements will be discussed briefly below.
Sampling Location. Sampling equipment must be fixed in a location that is easily and safely
accessible. If a primary control device is used, inflow and outflow samples should be taken a
short distance upstream of the device in a well-mixed section of such flow. If automatic
samplers are being used, the location of the intake tubing should not be significantly lower in
elevation than the autosampler unit. It was recommended that the difference between the intake
elevation and the sampler unit should be no more that 20 feet and preferably less (FHWA, 2001).
Studies have found that sample volumes became unacceptably variable at high lift heights
(FHWA, 2001). Also, in theory the representativeness of the collected sample with respect to
larger sediments may be compromised at high lift heights. Although some research shows that
this may not be as much of a problem as one may expect (FHWA, 2001). For BMP systems that
are arranged in a “treatment train”, sampling at intermediate locations within the system may be
useful to evaluate the significance of each individual component. Some special considerations
regarding selection of an in-stream sampling location are discussed in Section 4.4.
Sampling Method. Samples can be discrete, composite, or specialty. Discrete or grab samples
can be taken at a single point during the storm or sequentially throughout the storm. A discrete
sample taken at the onset of inflow can be used to characterize first-flush phenomenon
(GeoSyntec and ASCE, 2002). Discrete samples are required for some parameters like volatile
organic carbons that degrade if the sample is left exposed. Sequential discrete samples can be
used to create a “pollutograph” of pollutant concentration verses flow or time (depending on how
the series is weighted). A pollutograph can provide “insight into the performance of a BMP
[over time or] under various hydraulic loadings” (U.S. EPA, 1999). Peak concentrations and the
time during which water quality standards were exceeded (when applicable), also can be
determined from discrete sequential sampling (GeoSyntec and ASCE, 2002). Synchronizing the
times at which the discrete samples are pulled with flow measurements allows for accurate
calculation of loading rates and flow-weighted averages (i.e., event mean concentrations).
Composite sampling can be implemented in several ways. The simplest way, constant timeconstant volume composite, is not common and generally not acceptable for compliance with
4 - 22
stormwater regulations (FHWA, 2000). Time-weighted composites are not reliable for
estimation of mean storm concentrations or pollutant loads, and thus are not recommended for a
BMP monitoring program (GeoSyntec and ASCE, 2002). There are three flow-weighted ways to
collect composite samples that are considered suitable: (1) constant time-volume proportional to
flow rate; (2) constant time-volume proportional to flow-volume increment; and (3) constant
volume-time proportional to flow-volume increment (GeoSyntec and ASCE, 2002). For manual
sampling, the first option is most practical. For automatic samplers the third option is preferred
because it is generally more accurate than the other two flow-weighted sampling methods.
(GeoSyntec and ASCE, 2002). The increased accuracy is due to the combination of continuous
flow measurements (the first option is based on a single flow measurement) and the elimination
of sample splitting, which could introduce contamination (the varying volumes in second option
will require sample splitting).
The constant volume-time proportional sampling method adds individual aliquots to the
composite at a greater frequency when flows are high. Determining the exact autosampler
parameters, such as volume of aliquots and flow triggers, can be difficult. Low flows of smaller
storms can lead to insufficient sample volume, whereas, larger storms may fill the composite
sample bottle too quickly, truncating the later part of the storm. The resulting sample will not be
representative of the entire storm, thus calculation of an event mean concentration will be invalid
(Rushton, 2002). Problems may also arise when a large aliquot volume is combined with periods
of very high flows. Under these circumstances the flow-weighted sampler may be triggered to
collect the next sample while the previous sample is still being drawn. Thus, fixed autosampler
parameters are not appropriate for all storms. As a targeted storm approaches, estimates of storm
size and peak intensities should be made so that autosampler parameters may be adjusted as
necessary.
Sampling Frequency. The sampling frequency is the number of dry-weather samples taken or
storm events sampled. Given the great temporal and spacial variability of storm events, “a small
number of samples are not likely to provide reliable indication of stormwater quality at a given
site or the effect of a given BMP” (GeoSyntec and ASCE, 2002). Estimates of the sampling
frequency necessary to provide a reasonable average event mean concentration for a particular
location are eight to ten events (WEF and ASCE, 1998), however, statistical methods should be
used to determine an appropriate number of samples. These methods require an initial estimate
of sample variability, which can be obtained from existing data or from the literature, a minimum
level of detectable change, and a desired confidence level and power. Typically, calculations are
performed with a 95% confidence level at 80% power (GeoSyntec and ASCE, 2002). This
means that there is a 5% chance of finding significant change where none exists (type I error) and
a 20% chance of overlooking a significant change that does exist (type II error). The chosen
minimum level of detectable change makes a large difference in the necessary sampling
frequency. For example, the minimum number of samples required to detect a 5, 20, and 50%
change in mean EMCs for three key parameters at two monitoring locations are shown in Table
4-1.
4 - 23
Table 4-1. Minimum Sampling Frequency Necessary to Detect Significant Changes in
Mean Concentrations for Key Parameters at Two Sites in Portland, Oregon1
Site
1
2
Parameter
Minimum Number of Samples to Detect the Indicated
Percent Change2
5%
20%
50%
TSS
202
14
4
Phosphorus
224
16
4
Copper
442
29
6
TSS
61
5
2
Phosphorus
105
8
3
Copper
226
15
4
1
reproduced from (Strecker et al., 2000) pg. 183
80% power of detection in the mean of the EMCs
2
The increase in required sampling frequency between the three parameters indicates that
variability in percent reductions was much greater for copper compared to TSS and phosphorus
at each site. Their analysis shows that detection of a small change may take many years of
sampling every storm, given that some locations in the U.S., especially in the west/southwest,
experience only 10 to 20 measurable storms per year (Driscoll et al., 1989). Some have
estimated that at least five years are necessary to ensure important performance information is
not missed (Urbonas, 2000). The time of year during which these samples are taken is also
important. Due to the seasonality of many factors involved in BMP effectiveness, samples must
be collected covering all seasons, yet rarely is data collected in the winter in cold regions.
Irrespective of the above discussion, for the most representative analysis samples should be taken
as frequently as economically possible. Ideally, the sampling period should be continuous,
covering all seasons, in dry as well as wet weather. Synchronizing the samples with flow
measurements will produce accurate long-term pollutant mass reductions (Field and O'Connor,
1996).
Sample Representativeness. The NPDES permit requirements identify key characteristics that
result in a representative storm. The total precipitation and duration of a representative storm
should be within 50% of the average event for a given location, produce equal to or greater than
0.1 inches of precipitation, and have an antecedent dry period (i.e., less than 0.1 inches) of at
least 72 hours (GeoSyntec and ASCE, 2002). The antecedent dry period will set up a worst-case
scenario, allowing for build up of pollutants to be washed away with runoff. For BMP
monitoring purposes outside of a NPDES permitting application, strict adherence to the
4 - 24
representative storm is not necessary or likely not desirable. There is no truly “representative
storm” and monitoring only under these guidelines may skew the results. It is preferred that
monitoring be performed under a wide variety of conditions and storms (GeoSyntec and ASCE,
2002). Seasonality is another important factor of sample representativeness. Variation in human
activity, temperature, and precipitation type throughout the year should be considered when
determining the period over which the monitoring program is run.
During each storm event, sample representativeness can be determined using a percent capture
requirement (Caltrans, 2000). Percent capture is defined as the volume during which sampling
occurred divided by the total event volume. The California Department of Transportation
(Caltrans) (2000) also developed a minimum acceptable number of aliquots per composite
sample (when monitored on a flow-proportional basis) for an event based on the total
precipitation of the event in inches. Their percent capture and minimum number aliquots
requirements for a representative storm event are presented in Table 4-2. Once a sample has
been collected, this table is consulted and the sample is analyzed only if it is deemed
representative.
Table 4-2. Caltrans Requirements for Composite Sample Representativeness†
Total Event Precipitation
(inches)
Minimum Acceptable
Number of Aliquots
Percent Capture
Requirement
> 0.25
6
85
0.25 - 0.50
8
80
0.50 - 1.0
10
80
< 1.0
12
75
†reproduced from (Caltrans, 2000) pg. 10-10, Table 10-1.
The minimum acceptable number of aliquots for Caltrans shown in Table 4-2 are lower than that
reported elsewhere. In order to calculate an unbiased event mean concentration, WEF and ASCE
(1998) estimated between 15 and 30 aliquots must be taken per storm, while GeoSyntec and
ASCE (2002) estimated a lower range of between 12 and 16 aliquots per storm. No correlation
with event size was made for either of their estimates.
Sampling of dry-weather flows for baseline conditions should be less variable, but seasonal and
meteorological conditions should be considered. When the sampling location “is subjected to
base flows, runoff related flows should be separated from low flows” by at least 10% (FHWA,
2000).
4 - 25
4.5.2.5 Selection of Equipment and Materials
The two major types of equipment for a monitoring program are flow equipment and automatic
sampling equipment. A variety of secondary flow devices are available. GeoSyntec and ASCE
(2002) describes the different devices for measuring flow depth and considerations for their use.
Although not recommended, flow velocity equipment may be used with secondary devices to
determine flow volume in place of primary devices. If used, electromagnetic and ultrasonic
equipment for measuring flow velocity is preferred (GeoSyntec and ASCE, 2002).
Automatic samplers are usually well worth their capital expense because of the difficulty of
predicting weather. Paying a manual monitoring crew to be ready to start sampling as soon as
the runoff begins can result in wasted time and money (GeoSyntec and ASCE, 2002). However,
automatic samplers can be a significant source of variability. Shelley and Kirkpartrick (1975)
tested over 200 models of automatic samplers and found “marked differences in results obtained
with different types of equipment.” The type of pump used in the automatic sampler may affect
the representativeness of the sample. The intake or suction velocity achieved should equal or
exceed that of the flow being sampled (Shelley and Kirkpatrick, 1975). Larger sediments are
especially effected as lower capacity pumps may not be able to lift the larger solids from a lower
elevation intake into the sample bottles. Also, peristaltic pumps and sample transport lines with
internal constrictions and sharp twists and bends may break up or prohibit passage of larger
solids (Shelley and Kirkpatrick, 1975; Strecker, 1994). Because stormwater has a tendency to
become stratified, especially during low flows, in low velocities, or at the effluent end of a BMP
where larger and more dense particles have begun to settle out, use of a multi-leveled intake port
could provide a more representative sample (Field et al., 1997). Automatic sampler intake
tubing should have a minimum inner diameter of 1.0 to 1.3 cm to assure adequate capture of
larger particles (Shelley and Kirkpatrick, 1975). The intake tubing should be of opaque material
since clear plastic tubing may be subject to algae growth from exposure to sunlight, which will
contaminate the samples and may alter sample volume (Rushton, 2002). Also, steps may be
necessary to ensure that the intake tubing does not freeze during the winter months in northern
climates.
4.5.2.6 QA/QC Initiatives
QA/QC procedures ensure that the data collected is of reliable quality and, therefore, are an
important part of every BMP monitoring program. The QA/QC procedures provide a
quantitative measure of accuracy (FHWA, 2000). Field QA/QC initiatives include, field blanks,
field replicates, field sample volumes (e.g., percent capture as discussed in Section 4.5.2.4), and
chain of custody procedures that emphasize accurate labeling and other details such as legal
issues (GeoSyntec and ASCE, 2002). Laboratory QA/QC initiatives include, method blanks,
laboratory replicates, matrix spikes, and external reference standards (GeoSyntec and ASCE,
2002). Acceptable levels of variance and error should be established.
4 - 26
4.5.2.7 Quality Assurance Project Plan
The main product of the BMP monitoring program design phase is the QAPP. The QAPP lays
out all the design details discussed in Sections 4.5.2.1 through 4.5.2.6. It will serve as a guide for
all personnel involved. This will ensure consistency in technique between the program’s team
members. Thorough review, especially in terms of quality assurance, is highly recommended for
a QAPP before implementation. A good source of guidance regarding the preparation of a QAPP
is the U.S. EPA’s Guidance for Quality Assurance Project Plans (U.S. EPA, 2002b).
4.5.3 The Implem entation Phase
The implementation phase involves three main actions, equipment installation and testing,
sample handling and processing, and preliminary review of results (FHWA, 2000). The
preliminary review of results compares collected data against the QA/QC initiatives established
in the QAPP and checks that the data are reasonably close to expected ranges. All results that
fail the QA/QC measures should be flagged and eliminated if necessary. Operational differences
between the QAPP and actual practice inevitably occur during the implementation phase. All
changes to the QAPP should be documented.
4.5.4 The Evaluation Phase - Quantifying BMP Efficiency
Once a good set of quality assured data has been produced, the evaluation phase begins. The
evaluation phase has one main objective, data analysis. Sample processing typically produces
pollutant data in the form of concentrations. Pollutant loads can be calculated through the
mathematical combination of concentration data with the associated flow data. Loads are useful
information when evaluating long-term impacts (GeoSyntec and ASCE, 2002). Evaluation of
BMP efficiency over time within a single storm event on the basis of concentration or loading
requires the inflow and outflow data either to be matched in time or offset by the hydraulic
residence time of the BMP. Neither approach is considered acceptable. Generally BMPs, even
detention-type BMPs, do not act like a completely-stirred tank reactor (i.e., there is some lag
between the influent and the effluent concentration) nor do they act like an ideal plug flow
reactor (Law and Band, 1998). When it rains and a retention type BMP begins to produce
effluent, most or all of the water leaving the BMP is the water left in the BMP from the previous
storm. In fact, if a BMP was designed to hold a volume of water equal to that of the average
storm, as is often the case, 60 to70% of all storms would only displace the stagnant volume
without ever being a major part of the effluent (Strecker, 1994). Thus, matching influent and
effluent in time is not accurate. Offsetting the inflow and outflow by the hydraulic residence
time implies that the BMP acts as a perfect plug-flow reactor that is also generally not accurate
(GeoSyntec and ASCE, 2002). Thus, inner storm analysis is not recommended.
Under most circumstances the most useful parameter for comparing/combining pollutant data
between storms is the EMC (GeoSyntec and ASCE, 2002). Formally, the EMC is the total
pollutant load of the storm event divided by the total volume of the storm event. More simply,
4 - 27
the EMC is the concentration of a composite sample when flow-weighed sampling is performed.
The usefulness of the EMC diminishes when comparing pollutant data between BMPs. In these
cases, pollutant loads, not concentrations, are much more appropriate because loads are
normalized by volume. Both flow-weighted composite sampling and discrete sampling with
synchronized flow measurements can be used to calculate loads.
BMP pollutant removal efficiency is then calculated using any number of methods, including
percent removal, summation of loads, regression of loads, reduction in mean concentration,
irreducible concentration, achievable efficiency, removal relative to water quality limits, Lines of
Comparative Performance©, various multi-variate and non-linear models, effluent probability
method and linear regression of input versus output concentrations. The details of all but one of
these methods (linear regression of input versus output concentrations, discussed in detail below)
and their strengths and weaknesses are covered in GeoSyntec and ASCE (2002). The majority of
the eleven methods to evaluate BMP pollutant removal efficiency listed here are rarely used or
used only under specific circumstances. However, three of the methods (percent removal,
effluent probability method, and linear regression of input versus output concentrations) warrant
further discussion.
The most common method to evaluate the pollutant removal efficiency of a BMP is percent
removal. Sometimes referred to as the efficiency ratio, percent removal is defined as the average
difference between the inlet and outlet EMCs, divided by the average inlet EMC. This measure
of efficiency has been faulted because: (1) BMP percent removals are typically not uniform
across a wide range of influent water quality; and (2) a low influent EMC can result in a low
BMP effectiveness, even if the effluent concentration is near or at the limit of detection (Clar et
al., 2003). Current regulations or performance standards based on percent removals are ironic in
that those installations with the worst water quality are those where the percent removal
standards are most often met, whereas installations in areas with good upstream source controls
often produce failing percent removals despite their superior water quality (Urbonas, 2000).
Therefore, summary statistics on the inlet and outlet data should always be included in reports on
BMP effectiveness (Strecker et al., 2000). Despite its possible pitfalls, anticipated percent
removal is almost always used in selection and design of BMPs for pollution control. Therefore,
percent removal is a practical method of evaluating BMP monitoring data, especially when it can
be viewed along side particle settling velocity and/or size distribution for the same data set.
When care is taken to perform monitoring with a mass balance approach percent removal can be
a reliable measure of pollutant removal efficiency (Field and O'Connor, 1996). Long-term
continuous monitoring must be performed when using such an approach to overcome the
variability of influent and effluent characteristics with respect to dry- versus wet-weather flows,
season, and year.
The method recommended by GeoSyntec and ASCE (2002) for evaluating BMP effectiveness is
the effluent probability method. This method is statistically based. Statistical evaluation of the
observed differences is important in terms of establishing the BMP had an effect given the great
degree of variability inherent in stormwater monitoring (Strecker et al., 2000). The average
4 - 28
influent and effluent EMCs are examined for a statistically valid difference, then the data
(usually log-transformed) are plotted on a standard parallel probability plot. Along with an
indication of central tendency and variance, these plots will show useful information such as the
degree of consistency of removal over the full range of observations (GeoSyntec and ASCE,
2002). When using the probability method, care should be taken in averaging the influent and
effluent EMCs. The time-scale across which the data is aggregated should not be more than
seasonally based, due to the seasonality of stormwater and BMP environmental characteristics
(Law and Band, 1998). Such seasonal variability may mask statistically valid differences when
averaged annually. A limitation of the effluent probability method is the large data requirement
necessary to form a clear and useful picture of BMP performance efficiency. Presentation of
standard box-and-whisker plots to summarize the log-transformed inflow and outflow data also
have been recommended (Strecker et al., 2000). More recently, analysis of the data in the BMP
stormwater database showed the raw inflow and outflow data was much less variable that the
combination of this data into percent removals (Strecker et al., 2004). Average effluent water
quality was suggested as the most useful descriptor of BMP efficiency.
The most recent method for evaluating BMP effectiveness is linear regression of input versus
output concentrations proposed by Barrett (2004). This method pairs influent and effluent EMCs
on a storm-by-storm basis. The resulting linear equation can reveal several key points about
BMP performance. If the regression line is not statistically significant, then the influent water
quality does not have an effect on the effluent water quality for the particular BMP and parameter
in question. In this case, reporting the expected effluent water quality (i.e., average effluent
EMC) will be more useful. If the regression is significant, the constant of the regression can be
likened to the irreducible effluent concentration of the BMP. If the constant is near zero or
negative, then 100% efficiency may be achievable for the BMP and parameter in question. The
slope of the regression is an indication of percent removal. When the influent concentration is
sufficiently larger than the regression constant, then the expected percent removal will be
approximately one minus the slope. If the slope is greater than one, net export of the parameter
can be expected. The Barrett linear regression method can also be used to assess pollutant load
reductions (Barrett, 2004). Calculation of the load reductions requires water balance
measurements in order to determine the fraction of influent lost to infiltration and
evapotranspiration.
A possible disadvantage of the Barrett linear regression method is that variability between storms
in terms of flow, intensity, antecedent periods, etc. will translate into overly variable pairings
between influent and effluent concentrations. Therefore, it has been asserted that matching
average influent and effluent concentrations or loadings on a storm-by-storm basis may be of
limited value especially for those BMPs with permanent volumes (GeoSyntec and ASCE, 2002).
However, it is undeniable that in most cases the influent concentration will have some effect on
the effluent concentration. By assessing the entire data set as a whole, the Barrett linear
regression method has the advantage of allowing influent/effluent pairing while accounting for its
variability.
4 - 29
One final issue confronted during data analysis is how to handle non-detects. Non-detects are
measurements that are lower than the minimum detection limit specified by a particular method.
Traditionally data that fall below the limit of detection have been treated in one of four ways: (1)
excluded from the analysis altogether; (2) included in the analysis as zeros; (3) included in the
analysis as at the detection limit; and (4) included in the analysis as half the detection limit.
Using any one of these methods will bias the results. Option (1) will overestimate the mean,
while the remaining three options will underestimate the mean (Strecker, 1994). FHWA (2000)
suggested using option (4) only if the percentage of non-detects is less than 15% of the data. If
the percentage of non-detects is between 15 and 50%, “it is necessary to provide adjusted
estimates of central tendency and dispersion that account for data below the detection limit”
(FHWA, 2000). After reviewing all available techniques, Ruby and Kayhanian (2003)
developed a tool that imposes a log-normal probability distribution on the detected values, then
“assigns” expected values to the non-detects using regression and probability analysis. For data
sets with 50 to 90% non-detects, FHWA (2000) recommends changing the statistical analysis to a
percentile greater than the actual percentage of non-detects, while data sets above 90% nondetects do not provide enough useful data to perform any statistical analysis. However, even data
sets with greater than 90% non-detects still provides useful information and should be reported.
4.5.5 The Presumptive Approach
A BMP monitoring program is quite complex and involves significant effort and resources, as
can be seen from the above discussion. Consequently many municipalities cannot afford to
implement BMP monitoring studies. Yet at the very least, compliance with such rules as the
80% TSS removal rule for BMPs must be addressed where required. The presumptive approach
to monitoring uses existing monitoring data to develop an average removal percentage for similar
BMP types in similar geographical areas. As long as the BMP is designed, constructed, and
maintained properly, it is assumed that the average removal percentage is being equaled. NJDA
et al. (2000) has developed such a system for BMP monitoring in the state of New Jersey. They
composed a list of thirteen accepted BMPs and their expected average TSS percent removals.
The 80% TSS removal rule is assumed to be met as long as a BMP or a set of BMPs has been
properly installed such that the system has an expected average TSS percent removal equal to or
greater than 80% according to their list. Although the presumptive approach is simple and easily
applied, more good quality monitoring data is necessary before the average reported percent
removals in the literature are reliable enough for use in this manner.
REFERENCES
Argue, J.R. (1995). Towards a universal stormwater management practice for arid zone
residential developments. Water Science and Technology 31(1): 15-24.
Barrett, M.E. (2004). Performance comparison of structural stormwater BMPs. Water
Environment Research: awaiting publication.
4 - 30
Baxter, R. (2002). Grassroots stormwater programs: getting the public involved. Stormwater
3(6): 74-77.
Brick, C.M. and J.N. Moore. (1996). Diel variation of trace metals in the upper Clark Fork River,
Montana. Environmental Science and Technology 30(6): 1953 - 1960.
California Department of Transportation (Caltrans). (2000). Guidance manual: stormwater
monitoring protocols (second edition). CTSW-RT-00-005. Sacramento, CA.
http://www.dot.ca.gov/hq/env/stormwater/special/index.htm. May 2000. Accessed on
2003.
Church, P.E., G.E. Granato, and D.W. Owens. (1999). Basic requirements for collecting,
documenting, and reporting precipitation and stormwater-flow measurements. Report#
99-255. United States Department of the Interior, United States Geological Survey,
Northborough, Massachusetts. http://www.stormwaterresources.com/library/136PLUSGSMonitoring.pdf
Clar, M., B.J. Barfield, and S. Yu. (2003). Considerations in the design of treatment best
management practices (BMPs) to improve water quality. EPA 600/R-03/103. U.S.
Environmental Protection Agency, Office of Research and Development, Cincinnati, OH.
www.epa.gov/ORD/NRMRL/Pubs/600r03103/600r03103.htm October 2003.
Clary, J., J. Kelly, J. O'Brien, J. Jones, and M. Quigley. (2001). A Key Tool to help communities
meet Phase II stormwater requirements. Stormwater 2(2): 44-48.
Driscoll, E.D., G.E. Palhegyi, E.W. Strecker, and P.E. Shelley. (1989). Analysis of storm event
characteristics for selected rainfall gages throughout the united states - draft. U.S.
Environmental Protection Agency, Washington, DC. November 1989.
England, G. and B. Rushton. (2003). Monitoring guidelines for measuring stormwater gross
pollutants, http://www.stormwater-resources.com/library.htm#BMPs, (#204PL).
Stormwater News, Cocoa Beach, Florida.
Federal Highway Administration (FHWA). (2000). Stormwater best management practices in an
ultra-urban setting: selection and monitoring. FHWA-EP-00-002. U.S. Department of
Transportation, Federal Highway Administration, Washington, DC.
http://www.fhwa.dot.gov/environment/ultraurb/index.htm.
Administration, F.H. (FHWA). (2001). Guidance manual for monitoring highway runoff water
quality. FHWA-EP-01-022. U.S. Department of Transportation, Federal Highway
Administration, Office of Natural Environment, Washington, DC. June 2001.
Field, R. and T.P. O'Connor. (1996). Swirl technology: enhancement of design, evaluation, and
application. Journal of Environmental Engineering 122(8): 741-748.
Field, R., D. Averill, T.P. O'Connor, and P. Steel. (1997). Vortex separation technology. Water
Quality Research Journal of Canada 32(1): 185-214.
GeoSyntec Consultants Urban Drainage and Flood Control District and Urban Water Resources
Council of ASCE (GeoSyntec and ASCE). (2002). Urban stormwater BMP performance
monitoring: a guidance manual for meeting the national stormwater BMP database
requirements. EPA 821/B-02/001. U.S. Environmental Protection Agency, Office of
Water, Washington, DC.
4 - 31
Herricks, E.E. (1996). A protocol for wet weather discharge toxicity assessment. In: Water
Environment Foundation Technical Exposition and Conference, 69th Annual, Dallas, TX.
Water Environment Foundation, Pages 13 - 24.
Hillegass, J.B. (2003). Using an indicators database to measure stromwater program
effectiveness in Hampton Roads. In: Proceedings of National Conference on Urban
Stormwater: Enhancing programs at the local level, February 17 - 20, 2003, Chicago, IL.
U.S. EPA, EPA 625/C-03/003.
Kaiserli, A., D. Voutsa, and C. Samara. (2002). Phosphorus fractionation in lake sediments ­
Lakes Volvi and Koronia, N. Greece. Chemosphere 46: 1147 - 1155.
Law, N.L. and L.E. Band. (1998). Performance of urban stormwater best management practices,
http://www.unc.edu/depts/geog/them/projects/BMP.html. University of Carolina - Chapel
Hill, Department of Geology, Chapel Hill, NC.
Marsalek, J. and S. Kok. (2000). Foreward - Urban stormwater management for ecosystem
protection. Water Quality Research Journal of Canada 35(3): 313 - 314.
Medalie, L. (2000). USGS Factsheet 114-00: Monitoring the effectiveness of urban management
practices in improving water quality of Englesby Brook, Burlington, Vermont,
http://water.usgs.gov/pubs/FS/FS-114-00. United States Department of the Interior,
United States Geological Survey.
New Jersey Department of Agriculture, New Jersey Department of Community Affairs, New
Jersey Department of Environmental Protection, and New Jersey Department of
Transportation (NJDA et al.). (2000). DRAFT revised manual for New Jersey: best
management practices for control of nonpoint source pollution from stormwater (Fifth
draft). http://www.state.nj.us/dep/watershedmgt/bmpmanual.htm. May 3, 2000.
O'Shea, M.L. and R. Field. (1992). Detection and disinfection of pathogens in storm-generated
flows. Canadian Journal of Microbiology 38(4): 267 -276.
Parikh, P. (2003). A history and overview of the clean water act and its application to stormwater
runoff. National Risk Management Research Laboratory Seminar Series, November 6,
2003, Cincinnati, OH.
Pettersson, K. (2001). Phosphorus characteristics of settling and suspended particles in Lake
Erken. The Science of the Total Enivronment 266: 79 - 86.
Pitt, R., M. Lalor, R. Field, D.D. Adrian, and D. Barbe. (1993). Investigation of inappropriate
pollutant entries into storm drainage systems. EPA 600/R-92/238. U.S. Environmental
Protection Agency, Office of Research and Development, Cincinnati, OH. January 1993.
Pitt, R., R. Field, M. Lalor, and M. Brown. (1995). Urban stromwater toxic pollutants:
assessment, sources, and treatability. Water Environment Research 67(3): 260 - 275.
Reinelt, L.E. and R.R. Horner. (1995). Pollutant removal from stromwater runoff by palustrine
wetlands based on comprehensive budgets. Ecological Engineering 4(2): 77-97.
Ruby, A. and M. Kayhanian. (2003). Improving stormwater monitoring. Stormwater 4(6): 12 ­
30.
Rushton, B. (1998). Processes that affect stormwater pollution, http://www.stormwaterresources.com/library/001PPocesses.PDF. Stormwater News, Cocoa Beach, Florida.
Rushton, B. (2002). BMP monitoring: methods and evaluations, http://www.stormwaterresources.com/library.htm#BMPs (#147PL). Stormwater News, Cocoa Beach, Florida.
4 - 32
Shelley, P.E. and G.A. Kirkpatrick. (1975). An assessment of automatic sewer flow samplers.
EPA 600/2-75/065. U.S. Environmental Protection Agency, Office of Research and
Development, Cinncinati, OH. December 1975.
Strecker, E. (1994). Constituents and methods for assessing BMPs. In: Stormwater NPDES
related monitoring needs, August 7 - 12, 1994, New York, NY. The American Society of
Civil Engineers, Pages 329 - 348.
Strecker, E., M.M. Quigley, and B.R. Urbonas. (2000). Determining urban stormwater BMP
effectiveness. In: National Conference on Tools for Urban Water Resources Management
and Protection Proceedings, February 7 - 10, 2000, Chicago, IL. U.S. EPA (EPA 625/R00/001), http://www.epa.gov/ORD/WebPubs/nctuw/Strecker2.pdf.
Strecker, E.W. and B.R. Urbonas. (2001). Assessing receiving water effects of urban stormwater
BMPs. In: Linking stormwater BMP designs and performance to receving water impact
mitigation, August 19 - 24, 2001, Snowmass Village, CO. The American Society of Civil
Engineers, Pages 426 - 437.
Strecker, E.W., M.M. Quigley, B. Urbonas, and J. Jones. (2004). Analyses of the expanded
EPA/ASCE international BMP database and potential implications for BMP design. In:
Proceedings of the World Water and Environmental Resources Congress - Critical
Transitions in Water and Environemntal Resources Management, June 27 - July 1, 2004,
Salt Lake City, UT.
Tailor, A. and T. Wong. (2002a). Non-structural stormwater quality best management practices ­
a literature review of their use, value, cost, and evaluation. EPA Victoria Technical
Report 02/11. Cooperative Research Center for Catchment Hydrology, Victoria,
Australia. http://www.catchment.crc.org.au/pdfs/Technical200213.pdf December 2002.
Tailor, A. and T. Wong. (2002b). Non-structural stormwater quality best management practices
- an overview of their use, value, cost, and evaluation. EPA Victoria Technical Report
02/13. Cooperative Research Centre for Catchment Hydrology, Victoria, Australia.
http://www.catchment.crc.org.au/pdfs/technical200211.pdf December 2002.
United States Environmental Protection Agency (U.S. EPA). (1996). Monitoring guidance for
determining the effectiveness of nonpoint source controls. EPA 841/B-96/004. Office of
Water, Washington, DC. September 1997.
United States Environmental Protection Agency (U.S. EPA). (1999). Preliminary data summary
of urban storm water best management practices. EPA 821/R-99/012. Office of Water,
Washington, DC. August 1999.
United States Environmental Protection Agency (U.S. EPA). (2001). Techniques for tracking,
evaluating, and reporting the implementation of nonpoint source control measures ­
urban. EPA 841/B-00/007. Office of Water, Washington, DC. January 2001.
United States Environmental Protection Agency (U.S. EPA). (2002a). Environmental monitoring
and assessment program - research strategy. EPA 620/R-02/002. Office of Research and
Development, National Health and Environmental Effects Research Laboratory, Research
Triangle Park, N.C. July 2002.
United States Environmental Protection Agency (U.S. EPA). (2002b). Guidance for quality
assurance project plans. EPA/240/R-02/009. Office of Environmental Information,
Washington, DC. December 2002.
4 - 33
Urbonas, B. (2000). Assessment of stormwater best management practice effectiveness (chapter
7). In: Innovative urban wet-weather flow management systems, Field, R., J.P. Heaney,
and R. Pitt, (editors). Technomic Publishing Co., Inc., Lancaster, PA. Pages 255 - 300.
Water Environment Federation and American Society of Civil Engineers (WEF and ASCE).
(1998). Urban runoff quality management. 1-57278-039-8 and 0-7844-0174-8. Water
Environment Federation, Alexandria, VA.
4 - 34
5
Effective Use of BMPs in Stormwater
Management
Bethany Madge
CONTENTS
5.0 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 - 1
5.1 BMP Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 - 2
5.1.1 Regulatory Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 - 3
5.1.2 Site Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 - 7
5.1.3 Stormwater Quantity Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 - 16
5.1.4 Water Quality Performance - Pollutant Removal . . . . . . . . . . . . . . . . . . . . . . . 5 - 20
5.1.4.1 Factors Affecting Variability in Pollutant Removal by Structural
BMPs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 - 20
5.1.4.2 Approximate Pollutant Removals of Structural BMPs . . . . . . . . . . . 5 - 24
5.1.4.3 Approximate Pollutant Removals of Nonstructural BMPs . . . . . . . . 5 - 27
5.1.5 Maintenance, Reliability, and Cost Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 - 30
5.1.6 Environmental and Community Acceptance Factors . . . . . . . . . . . . . . . . . . . . 5 - 40
5.2 Structural BMP Placement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 - 41
5.3 Integrated Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 - 43
5.3.1 Integration of Structural BMPs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 - 43
5.3.2 Integration of Nonstructural BMPs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 - 44
5.3.3 Integrating Structural and Nonstructural BMPs . . . . . . . . . . . . . . . . . . . . . . . . 5 - 45
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 - 47
5.0 INTRODUCTION
The effective use of BMPs in managing stormwater runoff is dependant on many variables, e.g.,
design, construction, selection, placement, and maintenance. Proper BMP design and
construction are essential; however, BMP design issues have been discussed earlier in Chapter
3.0 and will not be repeated here. This chapter addresses proper BMP selection, placement, and
maintenance, which are also essential (U.S. EPA, 1999a). Section 5.1, BMP Selection, is written
from the perspective that a manager has a predetermined site on which to place or apply the most
appropriate BMP. Maintenance issues are also discussed in this section. Section 5.2, BMP
Placement, takes BMP selection to the next level by applying the selection process to a
watershed. In this section the manager begins only with a predetermined watershed that requires
5-1
protection. The BMP selection process must then be applied to all possible sites (or multiple
sites simultaneously) to ascertain the best possible options to achieve watershed protection. In a
more holistic management plan, the use of multiple BMPs in an integrated approach that targets
multiple pollutant types and forms can improve effectiveness over individual BMPs used alone.
Integrated approaches also are investigated in this chapter in Section 5.3.
The information presented is drawn from a literature review. It is intended to assist stormwater
managers in making educated decisions that will best fit their situation. However, there is a great
deal of variability involved in BMP applications. Even if all the guidelines and
recommendations in this chapter are followed, the expected outcomes are not guaranteed.
Adherence to the guidelines can only help prevent known pitfalls and identify issues that need
consideration to assure the best chances for a positive result.
5.1 BMP SELECTION
There are a large number of BMPs from which to choose. The U.S. EPA’s Menu of BMPs has
identified approximately 130 individual BMPs, categorized into the six minimum requirements
put forth by the National Pollutant Discharge Elimination System (NPDES) Stormwater Phase II
Regulations (U.S. EPA, 2003b). New BMPs, especially in the pollution prevention and public
education and participation categories, are being developed on a continual basis. In addition
there are some BMPs that are a combination of individual BMPs, i.e., low-impact development
techniques. Thus, selection of one or more BMPs appropriate for a particular situation may be a
difficult undertaking. Given the large number of choices, elimination of inappropriate or less
cost-effective BMPs through a series of sequential steps will lead to a much smaller list of the
most reasonable choices from which a final decision can be made. Six BMP selection steps have
been identified and are explained below in Sections 5.1.1 through 5.1.6. These steps include:
(1)
regulatory considerations,
(2)
site factors,
(3)
stormwater quantity issues,
(4)
water quality performance - pollutant removal,
(5)
cost, reliability, and maintenance issues, and
(6)
environmental and community acceptance factors.
Even with the steps detailed here as guidance, BMP selection is a complicated matter. Barraud et
al. (1999) developed a computer software program, DELTANOE, to assist stormwater managers
through the BMP selection process. With the input of some site-specific facts, their program can
produce a list of possible solutions, or when given a favored solution, it can produce information
on technical feasibility and possible scenarios of failure. As an added benefit, DELTANOE also
provides guidance on design, construction, and maintenance on each proposed solution. In early
2003, the U.S. EPA also initiated a project to develop a framework for the selection of BMPs for
successful stormwater management (Lai et al., 2003). The integrated stormwater decisionsupport framework (ISMDSF) is slated to use public domain models for hydraulic, hydrologic,
5-2
and water quality routing in spatial and temporal scales. The models will be combined with
watershed characteristics, BMP effectiveness, and cost data to provide watershed managers with
viable BMPs or combinations of BMPs to achieve the overall water quality objectives. A
geographical information system (GIS) will be used as a foundation to give the ISMDSF a
watershed perspective. Both structural and nonstructural BMPs are expected to be included. The
estimated completion of the initial phase, development of the overall framework and the
watershed component, is in early 2005 (Lai et al., 2003).
5.1.1 Regulatory Considerations
In 1999, the U.S. EPA published a ruling to expand the NPDES Stormwater Phase I Regulations
of 1990 (U.S. EPA, 2003c). Once only affecting a select group of industry and municipalities
with populations over 100,000, the new rules (NPDES Stormwater Phase II Regulations) apply to
all urbanized municipal separate storm sewer systems (MS4s), regardless of size. Coverage of
industrial activities, construction sites in particular, was also expanded in the second phase. The
rules require development of a stormwater management program for permitting of stormwater
outfalls which are administered by the state authority through the NPDES as of December 2002
(U.S. EPA, 2000). Elements from each of six minimum control measures (see Table 5-1) must
be included in the stormwater management program. Thus, BMP selection for municipalities
covered by the NPDES Stormwater Phase II Regulations, must consider options in each of the six
control measure areas. As mentioned above, the U.S. EPA has identified many BMPs for each
category. Table 5-1 lists the six minimum control measures and possible BMP options for
fulfilling each measure.
Table 5-1. BMP Options under the Six Minimum Control Measures Required by
U.S. EPA’s NPDES Stormwater Regulations: Phase I and II1
Minimum Control
Measure
Best Management Practice Options
Î Public
<Lawn & garden activities education
<W ater conservation practices for homeowners
<Education on & programs for proper disposal of household hazardous wastes
<Pet waste management education
<Trash management education
<Education/outreach for commercial activities
<Outreach programs for minority & disadvantaged communities & children
<Classroom education on stormwater
<Storm water e duc ationa l materia ls
<Low-impact development education
<Educational displays, pamphlets, booklets, & utility stuffers
<Using the media for community outreach on stormwater
<Prom otional giveawa ys to promo te stormwater aw areness
<Pollution prevention education for businesses
Education &
Outreach
5-3
Table 5-1. BMP Options under the Six Minimum Control Measures Required by
U.S. EPA’s NPDES Stormwater Regulations: Phase I and II1
Minimum Control
Measure
Best Management Practice Options
Ï Public
<Storm drain stenciling
<Stream cleanup & mo nitoring
<Volunteer monitoring
<Reforestation programs
<W etland plantings
<Adopt-A-Stream programs
<Rain barrel & cistern programs 2
<W atershed organization development
<Stakeholder meetings
<Attitude surveys
<Community hotlines
Involvement &
Participation
Ð Illicit Discharge
Detection &
Elimination
Ñ Construction
Site Stormwater
Runoff Control
<Failing septic systems
<Industrial/business connections
<Recreational sewage
<Sanitary sewer overflows
<Identifying illicit connections
<W astewater connections to the storm drain system
<Illegal dumping
<Non-stormwater discharges
Runoff Control
<Land grading
<Permanent diversions
<Preserving natural vegetation
<Construction entrances
<Check dams
<Filter berms
<Gra ss-lined channels
<Riprap
Sediment Co ntrol
<Tem porary d iversion dikes
<W ind fences & sand fences
<Brush b arriers
<Silt fences
<Sediment basins & ro ck dam s
<Sediment filters & sediment
chambers
<Sediment traps
<Storm drain inlet protection
Others
<Turf reinforcement mats
<Vegetative covers
5-4
Erosion Control
<Chemical stabilization
<Mulching
<Permanent seeding
<Sodding
<Soil roughening
<Geotextiles
<Gradient terraces
<Soil retention
<Temp orary slope drains
<Temp orary stream crossings
<Vegetated buffers
<Phase construction
<Construction sequencing
<Dust control
Good Housekeeping
<General construction site w aste
management
<Spill prevention & control plans
<Vehicle m aintenance & washing areas
<Contractor certification & inspector
training
<Construction reviewer
<BM P inspections & maintenance
<Mod el ordinances
Table 5-1. BMP Options under the Six Minimum Control Measures Required by
U.S. EPA’s NPDES Stormwater Regulations: Phase I and II1
Minimum Control
Measure
Best Management Practice Options
Ò Post-
Structural BMPs
<Dry extended-detention pond
<W et pond s
<Infiltration basins
<Infiltration trenchs
<Poro us pavem ents
<Bioretention
<Sand & o rganic media filters
<Stormw ater wetlands
<Grassed swales
<Vege tated filter strips
<Catchb asin/Catchba sin inserts
<In-line storage
<Manufactured products for
stormw ater inlet
<Gre enroofs 3
<Rain gardens 2
<Dry wells 2
Non structural BM Ps
<Alum injection
<Buffer zones
<Open space design
<Urb an fore stry
<Conse rvation easem ents
<Infrastructure planning
<Narro wer residential streets
<Eliminating curbs & gutters
<Green parking
<Alternative turnarounds
<Alternative pave rs
<BM P inspection & maintenance
<Ord inance s for po st-construction runoff
<Zoning
Others
<Mod ular treatment system
<Dynamic vortex separators 3
Sou rce Controls
<Pet waste collection
<Automobile maintenance
<Vehicle washing
<Illegal dump ing control
<Landscaping & lawn care
<Pest contro l
<Parking lot & street cleaning
<Roadway & bridge maintenance
<Septic system co ntrols
<Storm drain system cleaning
<Alternative discharge options
for chlorinated water
Materials management
<Alternative products
<Hazardous materials storage
<Road salt application & storage
<Spill response & prevention
<Used oil recycling
<Materials management
Others
<Airplane deicing fluid recovery
system
<Catchbasin cleaning
<Coverings
<Employee training
<Flow diversion
<Handling & disposal of resid uals
<Environmental effects from highway
ice & snow removal operations
<Internal reporting
<Materials inventory
<Preventative maintenance
<Record keep ing
<Spill prevention planning
<Stormwater contamination
assessment
<Visual inspections
Construction
Stormwater
Management in
New
Development &
Redevelopment
Ó Pollution
Prevention/
Good
Housekeeping
for Municipal
Operations
1
Taken from U.S. EPA’s menu of BMPs website (U.S. EPA, 2003b)
BMPs expanded out of the On-Lot Treatment listing from the U.S. EPA’s menu of BMPs
3
Additional BMPs not specifically listed in the U.S. EPA’s menu of BMPs
2
5-5
It has been estimated that pollutant load reductions of 25 to 40% will occur when permitted
NPDES stormwater management programs are fully implemented (Taylor and Wong, 2002a).
However, these estimates may be overly optimistic because municipalities often lack the
resources to enforce full implementation of their stormwater programs. For example, auditing of
construction sites subject to stormwater controls under one municipality found an average
compliance rate of 50% (Taylor and Wong, 2002a). Yet, compliance and subsequent pollutant
load reductions may improve over time as public awareness increases and attitudes and behaviors
change. For instance, another municipality reported a 30 to 40% compliance rate in the first few
years of the stormwater management program, but 90% compliance nearly a decade later (Taylor
and Wong, 2002a). Enforcement initiatives also increase compliance, but enforcement and
education both require a commitment of time and money.
The total maximum daily load (TMDL) program of the Clean Water Act may also play a role in
BMP selection for stormwater outfalls whose receiving waters are listed as impaired and thereby
subject to a TMDL. If the stormwater outfall is already regulated under a NPDES permit and a
TMDL is required for its receiving water, then a wasteload allocation must be assigned to the
outfall. The BMP or set of BMPs selected must then be able to achieve effluent loadings that
meet the wasteload allocated. Even if a stormwater outfall is not subject to a NPDES permit, it
may still be addressed in a TMDL under the load allocation component if it’s receiving water
requires a TMDL. See Wayland and Hanlon (2002) for more information on BMPs and how
they fit into the TMDL program.
The Coastal Zone Act Reauthorization Amendments (CZARA), Section 6217, is another
regulatory consideration. This section of CZARA outlines additional stormwater requirements
for coastal states. The part of CZARA that is the most influential is the new development
measure. This measure requires either post-construction total suspended solids (TSS) loadings
remain at pre-construction levels or at least an 80% reduction in annual TSS loading of storm
runoff to the receiving water must be achieved after the construction is over and the site has been
allowed time to stabilize (U.S. EPA, 2002a). For coastal areas this one requirement may be the
primary driver for BMP selection. Only a few structural BMPs listed under the 5th minimum
control measure in Table 5-1 are capable of consistently achieving 80% TSS reduction
individually (NJDA et al., 2000). However, the 80% removal may be reached by using more
than one BMP per site. More information about TSS removal efficiency is presented in Section
5.1.4.
The Endangered Species Act is a fourth federal regulation that may affect BMP selection.
Waterways designated as critical habitat areas of endangered animals may require extra levels of
protection by BMPs (Clar et al., 2003). For example, cold water species protection may rule out
BMPs with unvegetated permanent pools that are likely to contribute to temperature pollution,
unless the BMP can be designed offline with significant shading over open pool areas (U.S. EPA,
2002a). Other examples include protecting existing hydrology for both fish and amphibians.
State and local municipal regulations will also need to be reviewed before final BMP selection.
5-6
5.1.2 Site Factors
Site suitability is one of the key factors to successful BMP performance, especially for structural
BMPs (U.S. EPA, 1999a). Thus, particular attention must be paid to site suitability in BMP
selection. Site factors such as drainage area characteristics, climate and meteorological
characteristics, and physical factors at the location of BMP installation need to be considered
during BMP selection. The information presented in this section as well as the remaining 5.1
subsections relates primarily to the more common structural BMPs listed in Table 5-1 (namely
those BMPs covered in Chapter 3).
Drainage area characteristics include watershed size, land use (existing, at full build-out, and
other planned intermediate land use changes), and proximity to sensitive receiving waters. Land
use is often used as a selection factor under the assumption that a successful BMP in one
drainage area will be successful in a second area with similar land uses. Studies have shown a
correlation between land use and stormwater quality (GeoSyntec and ASCE, 2002; Selvakumar
and Borst, 2003). However, GeoSyntec and ASCE (2002) cautions that over reliance on this
correlation may not yield the desired result due to considerable variation in stormwater quality
and BMP performance even within the same land use category. Whether the BMP will serve
new development or be installed as a retrofit is important. For example, certain BMPs such as
swales typically are more practical for a new site that does not already have a curb and gutter
system. The overall directly-connected percent imperviousness of the drainage area can make a
large difference in BMP selection. Ultra-urban sites with little or no pervious areas are
particularly difficult to retrofit. BMPs such as disconnecting roof leaders may only lead to
flooding of streets and basements if not planned carefully. Limited available land in ultra-urban
sites generally eliminates most structural BMPs with large land requirements, (e.g., retention and
detention ponds). Pocket retention or detention ponds, or pocket wetlands are design variations
of their respective larger BMPs that may be used when land is limited. Pocket variety BMPs
require less of a footprint, but serve a much smaller drainage area. Sand filtration systems due to
their compact size and structured nature have been increasingly used in ultra-urban areas (U.S.
EPA, 1999a). In fact, filters can be installed as underground devices which keep their footprints
to a minimum. A summary of drainage area characteristics as they apply to the most commonly
used BMPs is presented in Table 5-2. This table represents generalized information that may not
be applicable in all situations.
5-7
Table 5-2. Generalized BMP Suitability for Relevant Drainage Area Characteristics1
BMP Type
Drainage Area
Size2
Land Use
New Development or
Retrofit
both, depends on land
availability
High Percent
Imperviousness Areas
Land Area
Requirements
Dry extended- 10-acre
detention pond minimum
widely applicable
yes, depends on land
availability
substantial,
moderate for
pocket dry ponds
Wet retention
pond
20 - 25-acre
minimum, 5­
acre maximum
for pocket wet
ponds
yes, depends on land
widely applicable,
both, depends on land
availability
ideal for parking lots availability, pocket wet
ponds may be more suitable
for retrofit situations
substantial,
moderate for
pocket wet ponds
Infiltration
trench
5-acre maximum not appropriate for
construction areas or
other areas with a
high solids source
which may cause
premature clogging
Porous
pavement
maximum ratio
of contributing
drainage area to
pavement area is
3:1
new development only, soil
compaction probable and
cost prohibitive in existing
developments
both, retrofit of large areas
parking lots,
driveways, residential may become cost
roads with no heavy prohibitive
vehicles and low
traffic volume
5-8
sometimes, soil
compaction under trench
possible and infiltrated
water may interfere with
existing infrastructure
minimal
yes, distributive nature of
porous pavement may
cope with low infiltration
rates, and subsurface
drainage systems may be
installed when infiltration
isn’t sufficient without
incresing land
requirements
not applicable,
product
substitution with
no additional
space
requirements
Table 5-2. Generalized BMP Suitability for Relevant Drainage Area Characteristics1
BMP Type
Drainage Area
Size2
Land Use
New Development or
Retrofit
ideal for parking lot both, retrofit of large areas
islands, and between can be cost prohibitive
roads and sidewalks,
also good for rooftop
runoff
Bioretention
2 - 5-acre
maximum
Sand or
organic media
filter
both
widely applicable,
10-acre
ideal for parking lots
maximum for
surface filters, 2­
acre maximum
for perimeter or
underground
filters
Stormwater
wetland
20 - 25-acre
minimum, 5­
acre maximum
for pocket
wetlands
widely applicable,
both, depends on land
ideal for parking lots availability, pocket
wetlands may be more
suitable for retrofit
situations
Grassed swale 5-acre maximum ideal for roadsides
(residential and
highways) and other
areas with sheet flow
runoff
both, although not usually
practical for retrofit in
existing curb and gutter
areas
5-9
High Percent
Imperviousness Areas
Land Area
Requirements
yes, ideally suited
minimal
yes, ideally suited
minimal to
moderate for
surface filters,
minimal for
perimeter or
underground
filters
yes, depends on land
availability
substantial,
moderate for
pocket wetlands
no, moderately sized and
decentralized land
requirements are usually
not available
moderate, but
variable
Table 5-2. Generalized BMP Suitability for Relevant Drainage Area Characteristics1
BMP Type
Drainage Area
Size2
Vegetated filter depends on
strip
design, approx. 1
acre per 580-ft
strip length
Land Use
New Development or
Retrofit
both, depends on land
ideal for roadsides
availability
(residential and
highways) and other
areas with sheet flow
runoff, practical for
very small parking
lots only
1
High Percent
Imperviousness Areas
no, impractical due to
large land requirements
relative to the treated
drainage area
Source of data: (NJDA et al., 2000; NYSDEC, 2001; U.S. EPA, 2002a; Clar et al., 2003; U.S. EPA, 2003b; ARC, 2001)
Th e num bers g iven are app roxim ate. Lo cal situatio ns may require large r or allo w sma ller area s.
2
5 - 10
Land Area
Requirements
moderate
Climate and meteorological characteristics that should be considered in BMP selection include
rainfall frequency, duration, intensity, climate (e.g., arid or cold), and evapotranspiration
potential. Infiltration BMPs such as infiltration trenches and filter strips may not have the
capacity to handle frequent (e.g., back-to-back) and/or intense storms. Cold climates with sub­
freezing temperatures at least part of the year, may present difficulty, although design
modifications can usually be applied (GeoSyntec and ASCE, 2002). Arid climates with high
average temperatures also may be problematic for certain BMPs. The performance of BMPs
which rely on availability of water, such as vegetative systems (e.g., filter strips and swales) and
those BMPs with open water (e.g., wetlands and retention ponds) can decrease without an
adequate water supply. Under such conditions, they may even become a public nuisance
(GeoSyntec and ASCE, 2002). Climatic considerations for most common structural BMPs are
presented in Table 5-3. Evapotranspiration potential is a related issue. Frost (2003) suggests
plotting rainfall versus evapotranspiration potential on a monthly basis. Areas in which
evapotranspiration exceeds rainfall for a predominant amount of the year will be good sites for
infiltration BMPs and nonstructural BMPs such as disconnecting impervious areas. The soils in
these areas will dry out quickly and thus will be able to absorb more rainfall.
Finally, physical characteristics of the BMP site location are important factors in BMP selection.
Soil characteristics such as permeability and erosiveness are some obvious examples. Infiltration
systems necessitate infiltration rates of 0.5 in./h or above (GeoSyntec and ASCE, 2002).
Infiltration systems also typically require at least 4 feet of soil, thus depth to bedrock or
groundwater is important. If groundwater is near the surface, the possibility for groundwater
contamination must be considered before selection of an infiltrating BMP. Site slope is another
factor. For example, gradients for swales “should be as close to zero as possible and should not
exceed 5%” (Frost, 2003). The specific geological feature, karst geology, is not appropriate for
many BMPs as the formation of sink holes are likely (NJDA et al., 2000). Finally, the hydraulic
head of a particular site, or the difference in elevation between the drainage area and the BMP
site itself is important as most BMPs operate by gravity flow through the system (NJDA et al.,
2000). Some structural BMPs are not affected by physical location. These BMPs are entirely
manmade and self contained systems (e.g., in-line storage, catchbasins and catchbasin inserts,
manufactured products for stormwater inlets, and rain barrels and cisterns). Along with the
climatic considerations, Table 5-3 also summarizes physical site considerations for most
common structural BMPs. As before, the recommendations/limitations listed in Table 5-3 are
generalized. Local conditions and engineering requirements may call for more or less
restrictions.
5 - 11
Table 5-3. Generalized BMP Suitability for Relevant Climate and Physical Site Characteristics1
BMP Type
Dry
extendeddetention
pond
Climate
modifications necessary
for both cold and arid
climates
Wet retention may not be suitable for
arid climates where
pond
permanent pool may be
difficult to maintain,
design variation needed
in cold climates
Infiltration
trench
modifications necessary
for both cold and arid
climates
Soils
Depth to Bedrock/
Water Table
Slope
maximum of
15%
upstream,
locally
relatively flat
maximum of
pond and ground
suitable for all soils,
15%
water
can
intersect
highly permeable soils
unless high levels of upstream,
may require an
locally
impermeable liner to contamination
relatively flat
expected, then at
maintain permanent
least 2 - 4 ft to water
pool
table
maximum of
maximum of 20% clay 2 - 5 ft minimum
from the bottom of 15%
and 40% silt/clay
the trench to bedrock upstream,
content, infiltration
locally
rates between 0.5 and or the seasonally
high water table, at relatively flat
3 in./h
least 100 ft from
down-gradient
drinking wells
suitable for all soils,
highly erodible soils
may require lined low
flow channel
high levels of
contamination
require at least 2 - 4
ft to water table,
otherwise can be less
5 - 12
Karst
Geology
Hydraulic
Head
impermeable 6 - 8 ft
liner required
impermeable 6 - 8 ft,
liner required 4 ft for
pocket wet
ponds
not practical
due to sink
hole
formation
1 - 4 ft
Table 5-3. Generalized BMP Suitability for Relevant Climate and Physical Site Characteristics1
BMP Type
Porous
pavement
Climate
not suitable for cold
climates where sand or
salt deicing is used or
where pavement is
subject to compaction
and wear and tear from
heavy snow plows,
freezing of infiltrating
runoff may cause frost
heave if base of stone
reservoir is not below
frost line
Soils
Depth to Bedrock/
Water Table
maximum of 20% clay
and 40% silt/clay
content, infiltration
rates between 0.5 and
3 in./h, otherwise use
with under pavement
storage
completely
2 - 5 ft minimum
from the bottom of flat
the basin to bedrock
or the seasonally
high water table, at
least 100 ft from
drinking wells
2 ft minimum to
Bioretention minor modifications may suitable for all soils,
water table
water percolates
be necessary for both
through manmade bed
cold and arid climates
Sand or
organic
media filter
Slope
2 ft minimum to
minor modifications may suitable for all soils,
water percolates
water table
be necessary for cold
through manmade bed
climates, limited
application in arid
climates where need to
irrigate may outweigh
benefit
5 - 13
Karst
Geology
not practical
due to sink
hole
formation
Hydraulic
Head
not
applicable
maximum of impermeable 2 - 5 ft
liner required
5% for both
upstream and
local
maximum of
6 - 10%
upstream,
locally
completely
flat
not an issue,
filters are
generally
concrete lined
5 - 8 ft for
surface
filters, 2 ft
minimum for
perimeter
filters
Table 5-3. Generalized BMP Suitability for Relevant Climate and Physical Site Characteristics1
BMP Type
Climate
Soils
Depth to Bedrock/
Water Table
Stormwater
wetland
not suitable for arid
climates where
permanent pool and
vegetation may be
difficult to maintain,
design variation needed
in cold climates
limited application in
arid climates where need
to irrigate may outweigh
benefit
suitable for all soils,
highly permeable soils
may require an
impermeable liner to
maintain permanent
pool and/or wetland
soil conditions
suitable for most soils
with some restrictions
on the most
impermeable soils, or
poor soils which
cannot sustain healthy
vegetative cover,
highly erodible soils
will require a flatter
slope and a highly
stabilizing grass
variety to reduce flow
velocity
pond and ground
water can intersect
unless high levels of
contamination
expected, then at
least 2 - 4 ft to water
table
usually 2 ft
minimum to water
table, except in wet
swale design
variation that
intersects with
groundwater
Grassed
swale
5 - 14
Slope
Karst
Geology
maximum of impermeable
liner required
8 - 15%
upstream,
locally
relatively flat
Hydraulic
Head
3 - 5 ft,
2 - 3 ft for
pocket
wetlands
impermeable 1 ft
1 - 2% (4 ­
liner required
6%
maximum)
along the
length of the
swale, can use
check dams to
adjust slopes
Table 5-3. Generalized BMP Suitability for Relevant Climate and Physical Site Characteristics1
BMP Type
Vegetated
filter strip
Climate
Soils
Depth to Bedrock/
Water Table
limited application in
arid climates where need
to irrigate may outweigh
benefit
not suitable for soils
with high clay content
or otherwise low
permeable soils, or
poor soils which
cannot sustain healthy
vegetative cover
2 - 4 ft minimum to
water table to ensure
filter strip does not
remain wet
1
Slope
2 - 10%
locally,
< 2% will
encourage
ponding,
> 6%
encourages
concentrated
flow
Source of data: (NJDA et al., 2000; NYSDEC, 2001; U.S. EPA, 2002a; Clar et al., 2003; U.S. EPA, 2003b)
5 - 15
Karst
Geology
not practical
due to sink
hole
formation
Hydraulic
Head
NA
5.1.3 Stormwater Quantity Issues
Stormwater quantity issues include quantity (flood and drainage) control, peak flow reduction,
groundwater recharge, and water reuse. Stormwater quantity control is the historic purpose of
BMPs, yet it is successfully achieved by very few. Therefore, if stormwater quantity control is an
objective, BMP selection will be more limited in options. Flood reduction has been the primary
purpose of extended-detention ponds, or dry ponds (U.S. EPA, 2003a). Retention ponds and
treatment wetlands also provide stormwater quantity control (NJDA et al., 2000). However,
since these BMPs are usually large regional facilities with large catchment areas they should be
designed for full build-out conditions. Otherwise, the storage area available for larger storm
flows may become too small which will eventually lead to increased downstream flooding.
Other BMPs that provide peak flow reduction, such as in-line storage, bioretention, infiltration
systems, swales, and greenroofs may help reduce flooding, but cannot be relied on as the sole
mechanism for stormwater quantity control in areas where flooding is problematic. However,
peak flow reduction itself is an important factor. The excess energy contained in peak runoff
flows accelerates stream bed erosion and degradation of habitat in the receiving water (U.S. EPA,
1999a). This excess energy can also mobilize pollutants leading to decreased water quality.
Both the maximum flows reached and the time that flows remain elevated above normal levels
are important in peak flow reduction.
In some areas groundwater recharge is an objective of BMPs. Recharge has a dual benefit. In
addition to recharging essential underground aquifers, recharge also reduces stormwater runoff
volume, reducing runoff pollutant loads, and thereby improving BMP effectiveness in preserving
surface water quality. As with quantity control, an objective of groundwater recharge has a
limited number of possible BMPs that will be effective at achieving this objective. Infiltration
systems such as infiltration basins, infiltration trenches, and porous pavement are obvious
choices. Less obvious choices for groundwater recharge include vegetated filter strips, swales,
dry wells, and rain gardens (NJDA et al., 2000). A review of the U.S. EPA/ASCE’s BMP
database found a reduction of stormwater volume of almost 30 percent for vegetative filter strips
and 30 percent for dry detention ponds (Strecker et al., 2004). These reductions are likely to
directly contribute to groundwater recharge. Other BMPs analyzed (wet retention ponds,
hydrodynamic devices, and wetlands) were not found to contribute significantly to groundwater
recharge (i.e., runoff volumes were not reduced by these BMPs) (Strecker et al., 2004).
In areas where freshwater sources are scarce, using stormwater as a water source for direct use
can be a great benefit. The state of Florida, for example, is actively encouraging stormwater
reuse (FDEP, 2003). Stormwater runoff can be collected and reused to meet irrigation needs for
land uses, such as commercial areas and golf courses, or used for toilet flushing or industrial
process waters. The water reuse option is generally reserved for retention pond type BMPs for
large scale operations (U.S. EPA, 2003a). Cisterns act in a similar fashion, as do rain barrels
albeit on a much smaller scale. In addition to the savings associated with using stormwater as a
resource, “stormwater reuse can help to maintain a more natural, pre-development hydrologic
balance in the watershed” (U.S. EPA, 1999a). Hydrologic balance is achieved through irrigation
5 - 16
by releasing the stormwater slowly and allowing evapotranspiration, infiltration, and
groundwater recharge, thereby restoring area aquifers to their natural levels.
Table 5-4 contains a summary of the ability of common structural BMPs to address the four
issues of stormwater quantity. Two structural BMPs (catchbasins and catchbasin inserts, and
manufactured products for stormwater inlets) have no influence on quantity issues.
5 - 17
Table 5-4. Generalized BMP Suitability for Water Quantity Control Issues1
BMP Type
Flood Control
Peak Flow Reduction
Groundwater Recharge
Water Reuse Options
high, with minor design
modifications, low for
common design practice
of 2-yr storm control
low to moderate, depends on
local soil permeability and
possible basin liner or low flow
channel, highest when pond is
completely inundated
no
high, with minor design
Wet retention high, ideally suited, but may
pond
increase downstream flooding modifications, lower for
common design practice
in some cases
of 2-yr storm control
none to low, if local soil
permeability is high enough for
significant recharge then the
basin must be lined inhibiting
recharge
yes, ideally suited
Dry
extendeddetention
pond
high, ideally suited, ~30%
reduction in total runoff
volume but may increase
downstream flooding in some
cases
Infiltration
trench
moderate
moderate, when
high, ideally suited
successfully designed
and maintained all runoff
flow will be diverted to
groundwater
Porous
pavement
variable,
low to moderate for
monolithic, moderate to high
for modular
high, when successfully
designed and maintained
all runoff flow will be
diverted to groundwater
Bioretention low to moderate, ~40%
volume reduction
high, ideally suited when sited
appropriately, approximately 70
- 80% of annual rain falling
directly onto porous pavement
surface area may exfiltrate to
groundwater
low on an individual
none, runoff collected in an
basis, scale generally too underdrain system
small to have a
significant impact
5 - 18
no
no
possible, filtered water
collected in underdrain
system could be
diverted to a cistern or
rain barrel
Table 5-4. Generalized BMP Suitability for Water Quantity Control Issues1
BMP Type
Flood Control
Peak Flow Reduction
Groundwater Recharge
Water Reuse Options
Sand or
organic
media filter
none to low
none to low
none, runoff collected in an
underdrain system, exfiltration
design modifications can be
incorporated to provide some
recharge
possible, filtered water
collected in underdrain
system could be
diverted to a cistern or
rain barrel
Stormwater
wetland
moderate to high, may increase
downstream flooding in some
cases; large fluctuations in
water level normally
associated with flood control
may permanently damage
wetland function
high, with minor design
modifications, low for
common design practice
of 2-yr storm control
none, may be some immediately
following installation, but build­
up of sediment at the bottom will
eventually stop infiltration
possible, but support of
wetlands vegetation
necessitates a minimum
acceptable permanent
volume
Grassed
swale
low to moderate, reported
reduction in total runoff
volumes = 6 - 30% over curb
and gutter system, and ~47%
influent end versus effluent
end
low, 2 - 6% reduction
over conventional curb
and gutter system
low to moderate when properly
constructed and maintained;
none for wet swales
no
Vegetated
filter strip
low to moderate, ~30 - 40%
reduction in total runoff
volume (influent end versus
effluent end)
low
moderate when properly
maintained
no
1
Source of data: (Schueler, 1987; WEF and ASCE, 1998; COV, 1999; NJDA et al., 2000; U.S. EPA, 2002a; Clar et al., 200 3; U .S. EP A, 20 03b ; Bar rett,
2004a; Strecker et al., 2004)
5 - 19
5.1.4 Water Quality Performance - Pollutant Removal
Pollutant removal has become one of the main objectives for using BMPs. Since NPDES began
permitting point sources of pollution under the Clean Water Act, the nation’s waters have
improved significantly (U.S. EPA, 2000). However, as point source controls improved, diffuse
sources (especially, stormwater runoff), have increased in relative significance and are now a
leading contributor to receiving water impairment. In particular, urbanized areas “export large
quantities of pollutants during storm events” (U.S. EPA, 1999a). The NPDES Stormwater
Regulations (Phase I and II) and CZARA were passed to address this problem. The NPDES
Stormwater Regulations require pollutant reduction to the maximum extent practicable, while
CZARA makes use of TSS as a pollutant indicator requiring 80% reduction in TSS whenever
annual TSS loadings cannot be maintained at or below pre-development levels (U.S. EPA,
1999a; U.S. EPA, 2000). These regulatory initiatives and the use of BMPs in TMDL
requirements have been the main drivers behind the focus on BMPs as pollutant removers.
Many studies of the pollutant removal efficiencies of BMPs can be found in the literature. The
majority of these data are presented in percent removals, but observed effluent quality ranges are
also sometimes reported. It must be stressed that due to the variable nature of BMPs,
numerically reported efficiencies are not easily transferable between sites, years, or even storms
at the same BMP. Therefore, selecting a BMP based on reported removal numbers may not
produce the desired result. Literature values for BMP pollutant removal efficiency should be
considered general estimates and too much emphasis on these numbers could be misleading.
Pollutant removal efficiency is site specific and highly variable between storm events even
within the same site. To achieve the desired water quality performance, it will be more
advantageous to develop an understanding of factors that are known to contribute to BMP
variability instead of focusing on specific numbers. Thus, the first portion of this section will
explore the factors that affect variability in BMP water quality performance. Only after these
factors have been discussed will approximate numerical efficiencies be presented. These first
two subsections will focus primarily on structural BMPs. The water quality of nonstructural
BMPs will be covered in the final subsection.
5.1.4.1 Factors Affecting Variability in Pollutant Removal by Structural BMPs
The factors that affect variability in structural BMP water quality performance can be divided
into two main categories: (1) factors that affect perceived or measured performance; and (2)
factors that affect true or actual performance. Factors affecting perceived performance are
governed by how the samples were taken, measured, and analyzed. These are the elements that
must be considered when developing or reviewing a BMP monitoring plan. Some major points
are reviewed here briefly. Chapter 4, BMP Monitoring should be consulted for more detail.
Factors affecting true or actual performance will be discussed here in more depth.
(1) Factors That Affect Perceived or Measured Performance
Many parameters must be satisfied to collect a true representative sample. A few examples of
these parameters are accurate flow measurement (and synchronization with sampling time for
5 - 20
discretes) to ensure accurate flow-weighted sampling, a large enough sampling period to capture
the entire storm/runoff event and identify the presence of dry weather flows, and proper selection
and placement of sampling equipment and intake lines. Once a sample is taken, analytical and
human error during pollutant measurement may introduce additional variability. Finally, once
the data has been generated, choices made during data analysis can have a significant impact on
the final reported pollutant removal efficiencies. There are many different methods available to
evaluate BMP pollutant removal efficiency. In one case study, four different methods of
evaluating pollutant removal efficiency were compared using the same set of data. The four
methods were percent removal, summation of loads, regression of loads, and average percent
removal of individual storms (GeoSyntec and ASCE, 2002). The calculated efficiency varied
greatly from -2 to 82%. When comparing across BMP types, the method chosen to analyze the
data can skew the overall results. For instance, methods used to evaluate BMPs in which
infiltration is the dominant pollutant removal mechanism should be based on loads, not
concentrations. Comparing the pollutant removal efficiency of an infiltration basin to a wet pond
on the basis of concentrations will likely favor the wet pond, even if the infiltration basin is
resulting in less pollutant loading to the receiving water.
Another factor that will alter the perceived performance is whether bypass volumes are included
in efficiency calculations. If a BMP is designed with a bypass or to treat only a portion of the
total runoff volume, whether or not the efficiency is calculated based on the whole system (i.e.,
entire runoff volume) or only on the portion of runoff actually captured and treated by the BMP,
will make a large difference in the reported numbers. In the past it has not always been clear in
these situations which method of data analysis was used.
(2) Factors Affecting True or Actual Performance
In addition to variability in measured performance, the true pollutant removal efficiency is highly
variable among sites utilizing the same BMP type, and even among storms within one BMP.
Four governing factors have been identified as the key elements in pollutant removal which
produce this variability: (1) active pollutant removal mechanisms; (2) BMP design
characteristics; (3) influent pollutant properties and concentrations; and (4) conditions within the
BMP. The components within each factor sometimes overlap or interact with each other
resulting in a complexity that is extremely difficult to model with simplistic tools like percent
removals. Thus, pooling of water quality performance data often results in meaningless numbers.
Active Pollutant Removal Mechanisms
As discussed in Chapter 2, mechanisms for pollutant removal in BMPs include sedimentation,
filtration, infiltration, sorption, phytoremediation, biological uptake, biological conversion,
floatation, and natural degradation (e.g., photolysis, hydrolysis, or volatilization) (U.S. EPA,
1999a). Each mechanism provides a different efficiency of removal for different types of
pollutants. For example, sedimentation typically will provide good removal for TSS, but will not
remove dissolved pollutants. The actual removal efficiency of an active pollutant removal
mechanism is dependant on all three of the remaining governing factors. As shown in Chapter 2,
Table 2-4, most structural BMPs incorporate more than one of these mechanisms. Thus, a
5 - 21
particular type of BMP will be appropriate for certain pollutants depending upon the dominant
removal mechanisms.
BMP Design Characteristics
The basic design of each type of structural BMP has many variations. Many of these variations
can affect the pollutant removal capacity of the BMP. The following list provides some
examples of prominent BMP design features that affect pollutant removal. This list is not meant
to be an exhaustive coverage of all necessary elements in BMP design. BMP design is discussed
in detail in Chapter 3.
•
In pond (both detention and retention) and wetland type BMPs residence time is a key. It
will dictate how much sediment will settle out of the water column before the effluent is
discharged.
•
Length to width ratio, depth, and total storage volume will also be important for
sedimentation efficiency in pond and wetland systems.
•
For BMPs with permanent pools, the volume of the permanent pool in relation to the
area’s average storm rainfall volume will determine the amount of treatment afforded.
Larger pool volumes will produce better and less variable effluent quality (Barrett, 2004b;
Strecker et al., 2004).
•
Total length and slopes are keys for a swale or vegetated filter strip. A longer length or
lower slope will promote a greater level of soil infiltration and filtration by the vegetation
within the BMP.
•
The presence, condition, and type of vegetation will affect the efficiency of
phytoremediation.
•
In a sand filter, the grain size of sand will determine the filter pore size, which in-turn
will affect the efficiency of filtration.
•
The type of media used in a media filter will affect the efficiency of sorption and
filtration.
Influent Pollutant Properties and Concentrations
The properties and concentrations of the influent pollutants to a stormwater BMP will affect the
efficiency of their removal. For example, if the size distribution of solids in the stormwater
influent is predominantly on the small end, then sedimentation and size exclusion filtration (as
found in sand filters) will be less efficient than if the size distribution is predominantly on the
larger end. Another example is the percentage of particle association of pollutants such as metals
and nutrients. For influents with a high percentage of particle association, more pollutant
removal mechanisms will be involved resulting in a higher overall removal efficiency. However,
for influents with little particle association, the efficiency of phytoremediation, sorption, and
other pollutant removal mechanisms that are more effective for dissolved constituents will
govern the overall pollutant removal.
Influent pollutant properties and concentrations are determined by three very important, but
secondary factors: (1) drainage area characteristics; (2) storm characteristics; and (3)
climatological factors. Drainage area characteristics include land use, soil characteristics, site
geology, and site topography. A different suite of pollutants will be expected depending upon the
5 - 22
contributing land uses present. For example, runoff from high density highways will likely
contain relatively high levels of metals (e.g., lead (Pb), zinc (Zn), iron, copper (Cu), cadmium,
chromium, nickel (Ni), and manganese) due to normal operation and frictional wear of vehicles
(FHWA, 2000). The characteristics of the soil, such as infiltration rates and sorptive capacity can
influence the concentration (by reducing the total runoff volume) and the particle association of
pollutants, respectively. The underlying geology of the area will dictate the characteristics of the
soil. Finally, site topography can enhance the erosional energy associated with stormwater
runoff. In areas with steeply sloping terrain, runoff will reach greater velocities and therefore be
more likely to erode and carry larger soil particles.
Storm characteristics include intensity and duration of the rainfall, the length of antecedent dry
period, and the characteristics of the rainfall itself. Storm intensity and duration will impact the
energy of the runoff, thereby effecting its pollutant carrying capacity. Storm intensity times
duration, i.e. total storm volume, will affect pollutant concentrations through dilution. Rainfall
duration (along with the antecedent dry period which will dictate how close the soil moisture
content is to it’s saturation point) will also affect the amount of infiltration attainable. This is
especially important in infiltrating BMPs such as infiltration trenches. The length of the
antecedent dry period, up to a point, can affect the influent pollutant concentrations due to a
phenomenon called build-up/wash-off (WEF and ASCE, 1998). In between storms, processes
such as atmospheric deposition, wind erosion, and vehicle leaking or wear of mechanical parts
will accumulate pollutants on land and street surfaces. Rainfall runoff will wash some portion of
the accumulated pollutants away. The longer the antecedent dry period, the more time pollutants
have to accumulate, and thus the higher pollutant concentrations are likely to be in the runoff.
However, wind erosion may also remove pollutants creating a maximum build-up level.
Sometimes, the rainfall itself can contribute to the pollutant loading. This is especially true if
nitrogen is a target pollutant as rainfall is known to contain high ammonia and nitrate levels
(Rushton, 1998).
Climatological factors include temperature, solar radiation, season, and wind. Temperature will
effect the influent pollutant properties by altering all kinetic reaction rates involved.
Atmospheric deposition and adsoption/desorption are examples of kinetic processes that may be
affected by changes in temperature. Solar radiation, especially on unshaded, dark roofing
material and blacktop paved surfaces, can increase the temperature of the runoff. Temperature
itself is considered a pollutant in the Clean Water Act (Parikh, 2003). Seasonal effects on
influent pollutant properties include the additional loading of de-icing chemicals, salts, and sand
used during the winter. These constituents will be present in abundance during snow melt
events. Influent total phosphorus concentration have been reported to be lower during the winter
months (Barrett, 2004b). Large loadings of organic matter, primarily decomposing leaves will
contribute high biological oxygen demand in the fall. Nutrient loadings may be seasonally high
during the spring and summer months, particularly if the catchment area is fertilized. Finally,
wind, as mentioned earlier, can contribute to soil erosion during dry times and also affects the
patterns of atmospheric deposition.
5 - 23
Conditions within the BMP
Conditions within the BMP that affect pollutant removal efficiency include dissolved oxygen
levels, sediment cation exchange capacity, thermal stratification, BMP vegetation characteristics,
climatological factors, possible groundwater intrusion, and BMP age and regularity of
maintenance. Dissolved oxygen levels, in addition to being significant for oxygen-dependant
aquatic organisms, also may control the concentrations of certain pollutants within a retention
(wet) type BMP. Low dissolved oxygen concentrations may cause phosphorus and the metals
cadmium, copper, iron, and zinc, to leach out of bottom sediments (Rushton, 1998). The cation
exchange capacity of the sediments of wetland and wet pond BMPs will affect the efficiency of
sorption as a pollution removal mechanism. Thermal stratification within a BMP with a
permanent pool will affect settleability of solids due to the establishment of density gradients.
This effect causes poor performance in the winter months, especially when there is ice cover
(Barrett, 2004b). The type and condition of vegetation within a BMP will determine the
efficiency of phytoremediation mechanisms. In grassed swales and vegetated filter strips,
filtration efficiency will also be affected by the type and condition of the vegetation. The
climatological factors of temperature, solar radiation, season, and wind will effect the conditions
within a BMP. Changes in temperature will cause changes in the kinetic rates of processes
occurring within a BMP, such as adsorption/desorption, biological uptake, biological conversion,
and natural degradation processes. Solar radiation on a quiescent permanent pool BMP will
contribute to temperature pollution. Solar radiation will also increase pollutant removals through
photolysis (Law and Band, 1998). Season will affect the vegetation within a BMP. During the
growing season, nutrient uptake by vegetation will be at its peak, but during the fall die-back,
vegetation will not uptake nutrients as readily and may even release some of the nutrients they
had previously taken up. Released nutrients may be in the same form or a different form from
what was originally taken up. Excessive wind can increase the dissolved oxygen in a pond or
wetland type BMP. In shallow BMPs such as wet swales, wetlands, and shallow ponds, wind
will also interfere with settling and may even resuspend previously settled solids. Possible
groundwater intrusion into a BMP can contribute additional dissolved pollutant loadings or lead
to dilution which can mask measured efficiencies when groundwater is not accounted for.
Finally, BMP age and regularity of maintenance contribute variations in BMP pollutant removal
efficiency. If not removed in a timely manner, decay of gross solids may contribute to pollutant
loadings. Also, accumulation of sediments within a BMP will decrease efficiency by creating
excessive bypassing or short circuiting in filters and pond-type BMPs, respectively.
5.1.4.2 Approximate Pollutant Removals of Structural BMPs
The pollutant removal of structural BMPs has been the subject of many studies. As stated
earlier, the majority of reported data is presented in percent removals. This method of evaluating
BMPs is widely criticized (see Section 4.5.4 for more information), however, a proven better
measure for evaluating BMP pollutant removal efficiencies is not available and/or widely
accepted (Urbonas, 2000; GeoSyntec and ASCE, 2002; U.S. EPA, 2002a; Clar et al., 2003).
Recently some have suggested that observed effluent quality is a more robust method for
characterizing BMP performance (Strecker et al., 2004). Therefore, the approximate pollutant
removals of the most common structural BMPs are presented in terms of percent removals in
5 - 24
Table 5-5 and observed effluent quality in Table 5-6 for a select group of pollutants. Prior to the
presentation of Tables 5-5 and 5-6, it must be stressed again that due to the variable nature of
BMPs, numerically reported efficiencies should be considered general estimates only.
Table 5-5. Approximate Pollutant Removal Efficiencies for Common Structural BMPs1
Percent Removal of Select Parameters
BMP Type
TSS
TP2
TN2
NOx2
Metals
61
19
31
9
26 - 54
68 ± 10
(-33 - 99)
55 ± 7
(12 - 91)
32 ± 11
(-12 - 85)
34 ± 21
(-85 - 97)
36 - 65
(-97 - 96)
Infiltration trench
75
60 - 70
55 - 60
--
85 - 90
Porous pavement
82 - 95
65
80 - 85
--
98 - 99
804
65 - 87
49
15 - 16
43 - 97
Sand or organic
media filter
66 - 95
4 - 51
44 - 47
-95 - 22
34 - 88
Stormwater wetland5
71 ± 35
56 ± 35
19 ± 29
40 ± 68
0 - 57
38 ± 31
(-100 - 99)
14 ± 23
(-100 - 99)
14 ± 41
(-100 - 99)
13 ± 31
(-100 - 99)
9 - 62
(-100 - 99)
54 - 84
-25 - 40
204
-27 - 20
-16 - 55
Dry extendeddetention pond
Wet retention pond3
Bioretention
Grassed swale3
Vegetated filter strip6
1
all data from (U.S. EPA, 2003b) unless otherwise noted
TP = total phophorus; TN = total nitrogen; NO x = nitrite + nitrate nitrogen
3
summary of 35 studies for wet ponds and 20 studies for swales presented individually (U.S. EPA, 2003b): mean
± 95% confidence level (minimum - maximum)
4
(ARC, 2001)
5
pond/wetland system
6
lower number corresponds to 75 ft buffer; upper number corresponds to a 150 ft buffer
2
Although the data in Table 5-5 was taken primarily from one source (i.e., U.S. EPA, 2003b), the
variability in percent removals is quite clear. This is especially true for wet retention ponds and
grassed swales. The U.S. EPA’s menu of BMPs (U.S. EPA, 2003b) listed numerous individual
studies for these two categories. The variability in reported percent removals was extreme. In
combining the listed data, the approximate percent removals had quite large confidence levels.
The data in Table 5-6 is also taken from the U.S. EPA’s menu of BMPs (U.S. EPA, 2003b) and
although it appears to be less variable, a large degree of uncertainty still exists in many instances.
Thus, the data in Tables 5-5 and 5-6 corroborate the point that caution is strongly advised when
applying collected pollutant removal efficiency data to a different site.
5 - 25
Table 5-6. Observed Effluent Quality Performance for Common Structural BMPs1
Effluent Quality of Select Parameters
BMP Type
TSS
(mg/L)
TP
(mg/L as P)
TN
(mg/L)
NOx
(mg/L as N)
Total Cu
(µg/L)
Total Pb
(µg/L)
Total Ni
(µg/L)
Total Zn
(µg/L)
Detention
pond
32 ± 12
(2.5 - 140)
0.32 ± 0.04
(0.02 - 0.86)
1.8 ± 0.88
(0.45 - 6.0)
0.28 ± 0.052
(0.25 - 0.60)
20 ± 2.6
(0.5 - 82)
27 ± 6.4
(1.3 - 200)
4.8 ± 0.82
(0.5 - 11)
110 ± 18
(0.70 - 610)
Retention
pond
24 ± 3.9
(0.03 - 250)
0.39 ± 0.15
(0.01 - 22)
1.1 ± 0.09
(0.10 - 4.1)
0.12 ± 0.03
(0.00 - 2.0)
11 ± 1.7
(0.13 - 130)
14 ± 2.5
(0.13 - 130)
4.0 ± 0.96
(0.11 - 23)
24 ± 2.9
(1 - 350)
Infiltration
trench
240 ± 2602
(120 - 420)
NI3
2.0 ± 0.762
(1.4 - 2.3)
0.57 ± 0.322,4
(0.36 - 0.73)
NI
26 ± 252
(12 - 42)
NI
90 ± 842
(50 - 150)
Porous
pavement
24 ± 202
(0.55 - 52)
0.03 ± 0.012
(0.02 - 0.04)
2.95
0.554,5
8.4 ± 172
(0.24 - 30)
14 ± 122
(0.91 - 33)
NI
11 ± 8.82
(1.5 - 26)
Sand/Media
filter
11 ± 7.4
(2.5 - 55)
0.24 ± 0.04
(0.002 - 2.3)
NI
0.26 ± 0.012
(0.25 - 0.34)
17 ± 2.7
(1.2 - 150)
8.4 ± 1.6
(1.0 - 110)
4.5 ± 8.7
(2.0 - 22)
120 ± 22
(1.0 - 960)
Wetland
Basin
22 ± 8.1
(0.14 - 730)
0.13 ± 0.02
(0.00 - 1.4)
2.0 ± 0.60
(0.01 - 51)
0.22 ± 0.10
(0.00 - 8.1)
3.9 ± 0.53
(0.50 - 16)
1.6 ± 0.29
(0.10 - 10)
NI
47 ± 13
(2.0 - 500)
Grassed
swale
26 ± 23
(5.0 - 56)
0.28 ± 0.08
(0.04 - 2.7)
0.62 ± 0.10
(0.06 - 2.7)
0.85 ± 0.86
(0.08 - 3.7)
6.5 ± 1.3
(0.30 - 56)
4.3 ± 1.1
(0.50 - 33)
NI
34 ± 3.7
(4.0 - 150)
Vegetated
filter strip
37 ± 472
(5.0 - 56)
0.91 ± 0.62
(0.15 - 9.3)
NI
NI
7.2 ± 1.7
(1.0 - 17)
15 ± 12
(1.0 - 150)
2.4 ± 0.762
(2.0 - 4.3)
44 ± 11
(3.0 - 150)
1
all data from the National Stormwater BMP Database (http://www.bmpdatabase.org/): mean ± 95% confidence level (minimum - maximum)
base d on one stu dy on ly
3
No Information
4
nitrate o nly
5
one d ata po int only
2
5 - 26
5.1.4.3 Approximate Pollutant Removals of Nonstructural BMPs
The nonstructural BMPs listed in Table 5-1, under Minimum Control Measure #5, will provide
pollutant removal through source reduction or pollution prevention measures. Public education
and participation measures foster better personal decisions on issues that affect stormwater runoff
quality. The success of source controls that depend on public buy-in such as pet waste
management, littering, lawn care and landscaping practices, vehicle washing practices, and used
oil and hazardous waste collection and recycling programs rely heavily on education and
participation of the community. Effectiveness of these essential, but indirect BMPs is difficult to
assess. Based on self-reported changes in behavior, intensive participation programs will be
more effective at reducing pollutant loadings, than passive education measures such as brochures
(Taylor and Wong, 2002a). Further, more memorable educational measures such as theater
productions about stormwater pollution issues and creative signage will be more likely to affect
change.
An additional nonstructural BMP of an effective stormwater management program is the
detection and elimination of illicit connections to stormwater drainage systems. Illicit
connections may include improper sanitary sewage connections, effluent from improperly
operating septic tanks, leaking tanks or pipes from industrial sites and wastewater from
commercial car wash and laundry facilities (Pitt et al., 1993). When present, these discharges
may result in significant pollutant loadings. For example, the Wayne County Department of
Public Health estimated that their illicit connection elimination program kept over 2,000 pounds
of suspended solids from entering the storm sewer system over a four year period (U.S. EPA,
2002b). Depending on the type of illicit connections identified, other pollutants that may be
reduced by eliminating these connections include oxygen demanding substances, bacteria and
other pathogens, hydrocarbons, and toxic chemicals. Since the existence and relative volume of
illicit connections are highly site specific, the effectiveness of this BMP is also highly site
specific. Pitt et al. (1993) and U.S. EPA (1999c) discuss methods of detecting illicit discharges
and appropriate corrective measures.
Construction site runoff management using erosion and sediment controls have been reported to
be very effective at preventing solids from leaving the construction site. The average suspended
solids load reduction by erosion controls such as preventative construction planning and phasing
and intensive site mulching and seeding has been estimated as approximately 85%, while
sediment controls such as sediment fences and sediment basins are reputed to achieve
approximately 60 to 70% (Taylor and Wong, 2002a).
Although, the pollution removal efficiencies of most pollution prevention controls cannot be
measured directly, some studies have produced estimates of the water quality control benefits of
these nonstructural BMPs. A summary of this information is presented in Table 5-7. Also, as
mentioned above, a number of these measures are dependant on personal behaviors, thus the
effectiveness of public education and outreach and public involvement and participation
initiatives will indirectly affect the actual realized benefits.
5 - 27
Table 5-7. Generalized Water Quality Performance of Select Nonstructural BMPs1
BMP Type
Qualitative Control Information
Estimated Percent Load
Reduction
Pet waste collection
potential load reductions for bacteria
none available
Vehicle washing
untreated car wash effluent is of poor quality (mean concentrations in mg/L
from one study were copper = 0.386, lead = 0.113, total nitrogen = 4.11,
total phosphorus = 0.32, TSS = 178, and zinc = 0.387); loading of these
pollutants may be reduced if vehicle washing is performed on a vegetated
area where effluents have an opportunity to infiltrate; removal could
approach 100% if local ordinances require washing at facilities where
effluents are drained to sanitary sewer lines
none available
Landscaping and
lawn care
choosing landscaping plants and designs with low maintenance regimes can
decrease runoff volume through reduced watering requirements and
reductions in nutrient loadings through reduced fertilization requirements;
lawn and garden fertilization is estimated to contribute 80 - 85% of
phosphorus loading in the spring
none available
Storm drain system
cleaning
removal of grass clippings and leaves from inlet structures can reduce
organic carbon and nutrient loadings; significance of impact will depend on
the land use of the drainage area
copper: 3 - 4%
Roadway and bridge
maintenance
preventative maintenance and prompt repairs of structural degradation will
reduce solids loadings to runoff
none available
Pest control
integrated pest management techniques can reduce pesticide and herbicide
usage by 75 - 87% and 85 - 90%, respectively; information on how this
translates to reduction in runoff loads is not available
none available
5 - 28
Table 5-7. Generalized Water Quality Performance of Select Nonstructural BMPs1
BMP Type
Qualitative Control Information
Estimated Percent Load
Reduction
Parking lot and
street cleaning
impact on organic carbon and nutrient loadings may be high during fall
season due to deciduous leaf removal; access to the curb must be maintained
using enforced parking restrictions for maximum effect; actual efficiency is
highly dependant on sweeping frequency; frequencies need to be
approximately once a week to maintain non-negligible effect
Industrial good
housekeeping
practices
improved materials handling and storage practices can reduce sediment and
chemical loadings; the data shown are the estimated effects of improved
industrial housekeeping measures (education of managers and implementa­
tion of an auditing program using a before/after study design), the estimated
largest measure taken was moving one of three industrial material stockpiles
in the catchment area from an uncovered to a covered area
removal of grass clippings and leaves from inlet structures can reduce
organic carbon and nutrient loadings; significance of impact will depend on
the land use of the drainage area
cadmium, copper,
and zinc: 45%
chemical oxygen demand
(COD): 34 - 45%
iron: 13 - 60%
lead: 5 - 48%
litter: 95 - 100%
organic nitrogen: 12 - 45%
total phosphorus: 9 - 28%
TSS: 37 - 50%
soluble phosphorus: 45%
copper: 42%
lead: 72%
total nitrogen: 40%
total phosphorus: 49%
TSS: 8%
zinc: 83%
COD, total phosphorus, total
kjeldahl nitrogen (TKN), and
zinc: 5 - 10%
lead and TSS: 10 - 25%
none available
Catchbasin cleaning
Used oil and
hazardous chemicals
collection/recycling
assuming that all household hazardous wastes and used oil would otherwise
be dumped down storm drains, tens to hundreds of tons of these materials
could be prevented from causing receiving water pollution
1
Source of data: (Taylor and Wong, 2002a)
5 - 29
5.1.5 Maintenance, Reliability, and Cost Issues
Costs, including capital, operation and maintenance, and time and effort requirements associated
with maintenance can be determining factors in the ultimate selection of an appropriate BMP.
Specific costing issues are covered in detail in Chapter 6 of this document and, therefore, will not
be repeated here. Thus, this section will focus on BMP reliability and maintenance regimes
necessary to sustain BMP effectiveness.
One strong factor in BMP selection is reliability of the BMP technology. Planning and
installation of a BMP can be costly and time consuming. To achieve the expected benefits and
avoid wasted efforts, the chosen BMP must be reasonably reliable. Levels and factors relating to
the reliability of the available BMP technologies are presented in Table 5-8.
The level of maintenance can vary significantly from one BMP to the next. Factors that affect
the required level of maintenance include, BMP type, visibility of the BMP, landscaping,
upstream conditions and safety issues (NJDA et al., 2000). If not maintained properly, the
reasons for selection of the particular BMP, such as it’s basic level of pollutant removal
performance and flood protection, will be lost. For instance, if sediment is allowed to
accumulate in the bottom of a retention pond, eventually the hydraulics within the pond will
change and short-circuiting and/or a loss in treatment capacity may occur, thereby reducing the
pollutant removal efficiency (Law and Band, 1998). The accumulated sediment also will
decrease the storage volume of the pond, which in turn will decrease the flood control value of
the pond.
Operational needs of BMPs include visual inspection, which should be done annually as a
minimum and after intense storms (NJDA et al., 2000). Inspection checklists will be BMP- and
site-specific. Some maintenance activities will be driven by these inspections. For example,
bank stabilization of emergency spillways should only occur if inspectors find that the bank is
beginning to show signs of erosion. Other examples include cleanup of vandalism, removal of
burrowing animals, and seeding or sodding in areas of dead or damaged ground cover. These
activities are usually unplanned and sporadic in nature and as such are not considered selection
related. Other maintenance activities are expected occurrences that require routine attention and
thus can be factored into the overall cost of a BMP (as seen in Chapter 6). These maintenance
activities are described for some commonly used BMPs in Table 5-8. In summary, a higher
maintenance burden can be expected with infiltration basins, infiltration trenches, porous
pavement, and stormwater wetlands. A medium maintenance burden can be expected with dry
extended retention ponds, wet ponds, sand and organic media filters, catchbasins and catchbasin
inserts, manufactured products for stormwater inlets, and modular treatment devices. A low
maintenance burden can be expected with bioretention, grassed swales, vegetated filter strips,
greenroofs, rain gardens, dry wells, and rain barrels/cisterns. Finally, in-line storage devices
carry little maintenance burden at all since they are designed to be self-cleaning. However, they
should be inspected regularly for malfunction and/or clogging.
5 - 30
Nonstructural BMPs (with a few exceptions, such as street and catchbasin cleaning) are not
generally thought of as having maintenance issues in the usual sense of the word, however, they
do require ongoing effort. Education and public participation programs must continually evolve
to remain productive and reach new audiences. Although genuine participation programs are
more effective than purely educational programs, they require a much greater time commitment.
Participation programs require the fostering of partnerships between agencies and community
stakeholder groups (Taylor and Wong, 2002a). To maintain these partnerships in proper working
order takes continuing time and effort.
5 - 31
Table 5-8. Generalized Structural BMP Reliability and Maintenance Requirements1
BMP Type
Dry
extendeddetention
pond
Reliability
Required Maintenance
Activities
moderate to
high; design
criteria well
established,
but wide
range in
performance
indicates
room for
improvement
1. mow side slopes and
remove litter
2. remove debris from inlet
and outlet structures and
repair damage if
necessary
3. repair and re-vegetate
eroded area
4. remove sediment from
the forebay
5. remove sediment across
the entire pond bottom
Recommended
Maintenance
Interval
monthly
monthly and/or
following large
storm events
annually
(if required)
5 - 7 yrs
25 - 50 yrs
5 - 32
Consequences of Failing to Perform Maintenance
degradation of aesthetics (actual frequency will depend
on climate)
inlet blockage - possible upstream flooding; outlet
blockage - extended ponding within BMP which may
encourage mosquito breeding
loss of treatment value and increase in effluent
sediment loading
decrease in maintenance interval for the more energy
and cost intensive maintenance activity #5
increased resuspension such that pond becomes
pollutant source and decrease in flood control volume
Table 5-8. Generalized Structural BMP Reliability and Maintenance Requirements1
BMP Type
Wet
retention
pond
Reliability
Required Maintenance
Activities
moderate to 1. mow side slopes and
remove litter and/or
high; design
hydrocarbon build-up
criteria well
established,
but wide
2. remove debris from inlet
range in
and outlet structures
performance 3. manage and harvest
indicates
wetland plants (if
room for
included in design) and
improvement
remove invasive species
4. remove sediment from
the forebay (after ~50%
loss of total capacity)
5. remove sediment across
the entire pond bottom
(after ~25% loss of
permanent pool volume)
6. repair inlet/outlet
structures and undercut
or eroded areas
Recommended
Maintenance
Interval
monthly
monthly
annually
(if required)
5 - 7 yrs
Consequences of Failing to Perform Maintenance
degradation of aesthetics; hydrocarbon build-up may
damage plants and pond microbiology (actual
frequency will depend on climate and contributing land
use)
inlet blockage - possible upstream flooding; outlet
blockage - local flooding
may be required in nutrient sensitive areas where
seasonal senescence can lead to significant annual
nutrient releases; invasive species may decrease
wildlife habitat
decrease in maintenance interval for the more energy
and cost intensive maintenance activity #5
20 - 50 yrs
increased resuspension such that pond becomes
pollutant source and decrease in flood control volume
infrequently (as
needed)
loss of treatment value; possible upstream of localized
flooding
5 - 33
Table 5-8. Generalized Structural BMP Reliability and Maintenance Requirements1
BMP Type
Infiltration
trench
Reliability
Required Maintenance
Activities
1. rake out and remove any
low to
accumulated sediment
moderate;
and debris from trench
reported
and inlets
failure rates
of about
2. replace top layers of pea
50%, actual
gravel, filter fabric, and
failure rates
topsoil
estimated to 3. total rehabilitation to
be higher
maintain a minimum of
2/3rds the design storage
capacity and a 72 hr
exfiltration rate
4. excavate trench walls to
expose clean soil
Recommended
Maintenance
Interval
monthly
Consequences of Failing to Perform Maintenance
degradation of aesthetics; infiltration will be impeded,
leading to loss of treatment value and possible flooding
as needed (when caking of top layers leading to excessive bypass
clogged)
decrease in infiltration rates leading to loss of treatment
upon failure
(typically within value, extended ponding which may encourage
mosquito breeding, and possible flooding
5 yrs)
upon failure
5 - 34
decrease in infiltration rates
Table 5-8. Generalized Structural BMP Reliability and Maintenance Requirements1
BMP Type
Porous
pavement
Reliability
Required Maintenance
Activities
Recommended
Maintenance
Interval
1. vacuum sweep (and high as needed
low to
(typically
pressure wash
moderate
monolithic) the surface, quarterly)
(modular
removing any obstructing
type); low
sediment and debris
(monolithic
type) which 2. stabilize adjacent areas annually
clogs much
(if required)
more rapidly 3. replace deteriorating or annually
(if required)
spalling modular
pavement
4. relieve localized clogging upon localized
failure
by drilling ½ in. holes
through pavement every
few feet within affected
spot
5. total replacement
upon complete
failure (1-2 yrs
for monolithic)
5 - 35
Consequences of Failing to Perform Maintenance
decrease in infiltration rates leading to loss of treatment
value and possible flooding
unstable adjacent areas will contribute excess sediment
loading, reducing effective life of pavement
breakdown of pavement will contribute to clogging and
premature deterioration of adjacent fully operational
pavement
loss of treatment value; localized flooding; decrease in
maintenance interval for the more energy and cost
intensive maintenance activity #5
complete loss of treatment value and possible flooding
Table 5-8. Generalized Structural BMP Reliability and Maintenance Requirements1
BMP Type
Reliability
Bioretention
no
information
Required Maintenance
Activities
Recommended
Maintenance
Interval
monthly
1. remove any litter and
debris
2. remove and replace dead biannually
or diseased vegetation
3. add fresh mulch
annually
4. remove and replace entire 2 - 3 yrs
mulch layer
5. repair of structural
infrequent (as
features (e.g.,
needed)
underdrains)
5 - 36
Consequences of Failing to Perform Maintenance
degradation of aesthetics
degradation of aesthetics, loss of treatment value
increased erosion due to exposed soil
loss of treatment value due to breakdown of mulch
material and decrease in ion-exchange capacity
localized flooding; lack of drainage may kill plant root
systems
Table 5-8. Generalized Structural BMP Reliability and Maintenance Requirements1
BMP Type
Sand or
organic
media filter
Reliability
Required Maintenance
Activities
1. remove debris/floatables
low to
from inlet/outlet
moderate;
structures and sedimenta­
organic
tion chamber (if present)
media filters
such as peat,
peat/sand
2. remove sediment from
mix, and
sedimentation chamber
compost/
(if present)
sand mix
3. renew or replace top few
tend to clog
inches of filter bed
faster than
surface
sand alone
4. renew or replace entire
filter bed (once draw
down exceeds 24 hrs)
5. repair leaks or
deterioration of structural
components
Recommended
Maintenance
Interval
Consequences of Failing to Perform Maintenance
monthly
inlet blockage - possible upstream flooding; outlet
blockage - extended ponding within BMP which may
encourage mosquito breeding; loss of treatment value if
filter is bypassed due to blockage at either inlet or
outlet
as needed (when inadequate sedimentation will lead to accelerated filter
50% full)
clogging
annually (or as
needed)
filter bed sealing; excessive bypass volumes; loss of
treatment value
3 - 5 yrs
filter bed clogging; excessive bypass volumes; loss of
treatment value; and extended ponding
infrequent (as
needed)
loss of treatment value; possible localized flooding
5 - 37
Table 5-8. Generalized Structural BMP Reliability and Maintenance Requirements1
BMP Type
Stormwater
wetland
Reliability
Required Maintenance
Activities
moderate to 1. mow side slopes and
remove litter
high; design
criteria well 2. remove debris from inlet
established,
and outlet structures
but wide
3. remove and replace dead
range in
or diseased vegetation
performance
indicates
4. harvest and replace
room for
wetland plants
improvement
Recommended
Consequences of Failing to Perform Maintenance
Maintenance
Interval
monthly
degradation of aesthetics (actual frequency will depend
on climate)
monthly
inlet blockage - possible upstream flooding; outlet
blockage - localized flooding
annually
loss of treatment value; degradation of aesthetics;
(if required)
coverage of $ 50% of wetland surface area is
recommended
annually
may be required in nutrient sensitive areas where
(if required)
seasonal senescence can lead to significant annual
nutrient releases, however benefits of regular harvesting
are not well substantiated by field data
annually
loss of wildlife habitat
(if required)
5 - 7 yrs
decrease in maintenance interval for the more energy
and cost intensive maintenance activity #7
5. remove and replace
invasive vegetation
6. remove sediment from
the forebay (after ~50%
loss of total forebay
capacity)
7. remove sediment across 20 - 50 yrs
the entire pond bottom
(after ~25% loss of
wetland storage volume)
8. repair inlet/outlet
structures and undercut
or eroded areas
infrequently (as
needed)
5 - 38
decrease in flood control volume, increase in
resuspension potential such that wetland may become
pollutant source, vegetation may be choked out, and
wetland may become overly eutrophic leading to
aesthetic and odor problems
loss of treatment value; possible upstream of localized
flooding
Table 5-8. Generalized Structural BMP Reliability and Maintenance Requirements1
BMP Type
Grassed
swale
Vegetated
filter strip
Reliability
Required Maintenance
Activities
1. mow side slopes and
moderate;
remove litter and grass
subject to
clippings
disruption by
homeowners 2. remove and replace dead
or diseased vegetation
3. aerate and dethatch grass
on swale bottom
Recommended
Maintenance
Interval
monthly
annually
(if required)
annually (once
draw down takes
> 48 hrs)
5 yrs (or as
4. remove sediment
accumulation from swale needed)
bottom once treatment
volume reaches 25% of
design volume
1. mow grass and remove monthly
moderate;
concentration litter and grass clippings
of flows can
often lead to 2. remove and replace dead annually
erosion and
or diseased vegetation
(if needed)
failure
3. repair (reseed or resod) annually
any rills, gullies, or bare (if needed)
areas
infrequent (as
4. rototill or cultivate
surface of sand/soil bed needed)
1
Consequences of Failing to Perform Maintenance
degradation of aesthetics (actual frequency will depend
on climate); maintain height at 3 - 6 in. to retain
filtration value
loss of treatment value; degradation of aesthetics
decrease in infiltration rates leading to loss of treatment
value, extended ponding which may encourage
mosquito breeding, and possible flooding
decrease in flood control volume and decrease in
infiltration rates
degradation of aesthetics (actual frequency will depend
on climate); maintain height at 2 - 4 in. to retain
filtration value
loss of treatment value; degradation of aesthetics
loss of treatment value due to short-circuiting
decrease in infiltration rates leading to loss of treatment
value
Source of data: (NJDA et al., 2000; Urbonas, 2000; ARC, 2001; NYSDE C, 2001; U.S. EPA, 2002a; U.S. EPA, 2003b)
5 - 39
5.1.6 Environmental and Community Acceptance Factors
The final group of factors to consider before definitive selection of a BMP is environmental and
community acceptance factors. Environmental factors include the possibility of mosquito
breeding and providing wildlife habitat. Detention (dry) ponds, swales, stormwater wetlands,
and infiltration BMPs have the potential to become mosquito breeding grounds, especially if they
are not properly maintained. Before BMP installation, consider performing an assessment of the
local wildlife and their habitat needs. This may be more important in areas of new development
where the impact on available habitat is more pronounced. Natural, unmowed vegetation found
in stormwater wetlands, some swales, and vegetated wet ponds, can provide habitat for song
birds and small animals, whereas cleared and manicured BMPs, like detention (dry) ponds and
infiltration basins provide no wildlife habitat value (NJDA et al., 2000).
The two key community acceptance factors are: (1) safety issues; and (2) aesthetics, especially as
they relate to financial implications. Community acceptance is especially important if public
participation is required. Safety issues with BMPs are rare and those that do exist can usually be
resolved through good design. An example of this is installation of wire grates on outlet
structures to prevent sudden storm surges from sweeping someone into a storm drain. Another
safety concern is the drowning hazard posed by permanent volume BMPs (wet ponds and
wetlands). This can be reconciled by installing a fence around the perimeter of the BMP,
however, this will make maintenance more difficult and also detract from the aesthetic value of
these BMPs. Thus, community acceptance of permanent volume BMPs may be difficult to attain
in areas with a high density of young children.
Aesthetic issues are of particular importance for community acceptance, not only because the
property owners have to live side-by-side with the BMP, but also due to the financial
implications aesthetic value brings. BMPs that are aesthetically pleasing will appreciate the
value of the surrounding properties, and vice-versa. Thus aesthetics are often a key factor in
gaining community acceptance. Although aesthetics are usually a matter of personal preference,
a few aesthetic issues are more generally accepted. Odor nuisances caused by failing infiltration
type BMPs and over eutrophication of wet ponds and wetlands can be problematic. Cattails and
other emergent wetland vegetation in wet ponds, wetlands, and some swales are considered
unsightly to some people (NJDA et al., 2000) while not to many others. BMPs such as
infiltration basins and trenches and rain gardens, which are intended to create ponding areas that
drain over a relatively short period of time, can be either objectionable “soggy spots” or desirable
scenery depending on the viewer. Dry detention basins, especially if not well maintained, often
become large garbage collectors. On the other hand, BMPs like wet ponds, stormwater wetlands,
bioretention areas, rain gardens, and greenroofs can give a community an up-scale look, adding
to the aesthetic value of a property rather than detracting from it.
Public acceptance is paramount for BMPs that require public participation, such as rain barrels
and dry wells, or nonstructural BMPs like car washing and pet waste management ordinances.
Studies have shown that “the magnitude of induced behavioral change is likely to vary greatly on
the nature of the activity promoted and the target audience” (Taylor and Wong, 2002a). As an
5 - 40
example, Taylor and Wong (2002a) contend that an educational campaign to get people to wash
their cars on their lawns would have a greater level of public acceptance in a low-density
residential area than one that encourages the use of sewered wash bays. If one can attach a
personal benefit to a BMP that requires public participation, then public acceptance will be
higher and adoption of the BMP will occur at a much greater rate (Taylor and Wong, 2002a).
For instance, playing up the costs savings on water bills gained from using a rain barrel for
landscape watering needs will improve public acceptance of this BMP.
5.2 STRUCTURAL BMP PLACEMENT
BMP placement refers to the spatial scale covered by a single BMP or set of BMPs used to
protect the watershed as a whole. Three spatial scales of BMP placement have been identified;
onsite, sub-regional, and regional (Lai et al., 2003). The onsite spatial scale is defined as 10 to
100 acres, a size typical of a single commercial/industrial building lot or residential
neighborhood. The sub-regional scale is defined as 100 to 5,000 acres and represents the
township level. The regional scale has a drainage area greater than 5,000 acres, exemplifying an
entire watershed, sewershed, or county. There is currently no comprehensive way to make the
decision between a collection of small onsite BMPs located throughout the watershed or one
large regional control located downstream or a combination of the two, although research in this
area is underway (Sullivan et al., 2003). The integrated stormwater management decisionsupport framework, discussed earlier in Section 5.1, is to include BMP placement as a central
part of its overall BMP selection tool. Issues such as retrofit versus new development, cost, land
availability, and hydrologic/hydraulic/water quality routing principles will be factored into BMP
placement decisions (Lai et al., 2003). However, this project is still in its initial phase. In the
meantime, some general guidelines on BMP placement are available.
The issue of BMP placement is largely affected by whether the watershed in question is an
existing urban area or a new development. Retrofitting BMPs into an existing urban area places
many constraints on BMP placement. Existing drainage systems, limited available space, and the
high cost of land in these areas make applying large, regional BMPs much more difficult in
existing developments (Sieker and Klein, 1998). However, in some circumstances these
facilities have been shown to be cost effective (Strecker et al., 2002). Densely populated areas
are the most complicated to retrofit. The high percent imperviousness of such areas requires
rapid conveyance of the stormwater in order to avoid flooding (U.S. EPA, 1999a). Thus, the
question of BMP placement within high density areas is limited to source control techniques,
such as greenroofs, in-line storage management techniques in existing conveyance systems, or
retrofit of existing flood conveyance or detention basins. In existing lower density areas, some
onsite BMPs become feasible, such as use of rain barrels and dry wells. The expensive retrofit of
replacing curb and gutter systems with swales can also be used in existing low density areas.
Sand and media filtration systems are one regional control option that is often used in retrofit
situations due to its limited space requirements (U.S. EPA, 1999a). Sand filtration can even be
constructed underground thus requiring a very minimal footprint. However, larger BMPs
traditionally used as regional controls, such as dry and wet ponds, are usually not an option.
5 - 41
Despite these constraints, areas as densely populated as New York City’s Staten Island have
instituted a wide-reaching BMP program that successfully addressed placement issues (Vokral et
al., 2001). The New York City Department of Environmental Protection (NYCDEP) developed
the Staten Island Bluebelt Program to combat flooding problems. A knowledge of flood prone
areas was combined with a review of the regional topography, the locations of existing storm
sewers, and the ecological condition and hydrological patterns of local streams. This assessment
assisted the NYCDEP in the placement of BMP retrofits. Once high risk areas were identified,
property in the area was acquired and a system of small “pocket wetlands” were installed in areas
as small as single lots.
In new developments, the use of BMPs are more of a proactive approach rather than a reactive
approach. Thus, the issue of proper placement of BMPs in the watershed is more of a conscious
decision rather than a reaction to existing consequences as with retrofit situations. Onsite
practices can be advantageous because they combat the problem close to its source. This makes
onsite placement a good option for approximating pre-development hydrology. The onsite scale
is most often used due to its ease of implementation (Clar et al., 2003). Due to the small area
involved, onsite placement can reduce political or institutional complications. Many
municipalities have regulations requiring the use of onsite BMPs as part of their permits for new
construction. This has the benefit of shifting financial responsibility to the developer. Yet,
onsite placement is not without its problems. The decentralized ownership of onsite BMPs leads
to a wide range of BMP design, construction, and operation and maintenance practice issues
(Finnemore, 1982). In addition, options to treat existing runoff together with that from new
development could be missed. Under these circumstances, water quality and quantity control
benefits are difficult to assess and may be unreliable. Also, onsite placement of BMPs is
sometimes difficult to coordinate in order to reap real stream protection benefits on a watershed
scale.
Sub-regional and regional BMP placement options can be advantageous for several reasons (Clar
et al., 2003). Their affect on the receiving water is more easily observed and quantified. For
watershed management programs under the control of a public agency, sub-regional or regional
placement can be more cost-effective because one large BMP will cost less to build and maintain
than many smaller BMPs. Maintenance programs for BMPs at this level are more likely to be
consistent and long-term. And finally, the size of regional BMPs is conducive for recreational,
aesthetic, and wildlife habitat benefits that may not necessarily accompany smaller onsite BMPs.
Disadvantages to the wider scales of BMP placement include: (1) the need for comprehensive;
advanced planning; (2) large initial capital costs; (3) failure to provide significant flood relief; (4)
failure to provide downstream channel protection through peak discharge control; (5) blockage of
fish migration routes; (6) possible thermal pollution; and (7) accelerated erosion and water
quality degradation in the unprotected upstream feeder streams (Urbonas, 2000; Clar et al.,
2003).
5 - 42
5.3 INTEGRATED APPROACH
An integrated approach to stormwater management is considered more effective than any single
BMP alone. Case studies recognize that in-stream biological condition may not be measurably
protected or improved by an isolated BMP (Stribling et al., 2001). Stream degradation is rarely
caused by a single stressor and more than one type of BMP may be necessary to relieve all
contributing stressors. The NPDES Stormwater Phase I and II Regulations mandate an integrated
approach by requiring elements from each of the six minimum control measures shown in Table
5-1. Thus, evaluating the performance of an individual BMP may misrepresent the role of that
BMP in the overall management scheme (Taylor and Wong, 2002a). Integration of BMPs may
occur on three different levels: integration of more than one structural BMP, integration of more
than one nonstructural BMP, and finally integration of structural and nonstructural BMPs. The
following sections review each of these three levels.
The integrated approach discussed in this section is the integration of individual BMPs with each
other. However, a holistic integrated approach also should consider BMP integration within the
community. For example, incorporating small design features into the larger structural BMPs
can turn them into public amenities (Roesner, 1999). Parks and walking paths around
stormwater management facilities can add aesthetic and recreational value, turning BMPs from
community nuisances into community assets.
Finally, it is important to acknowledge that measured pollutant removal efficiencies (specifically,
percent removals) of individual BMPs may not be representative of the overall benefit to the
receiving water when the BMP is part of an integrated approach (Rushton, 2002). When
stormwater runoff is pre-treated by upstream BMPs, the influent water quality will be cleaner and
thus a large difference between influent and effluent pollutant concentration will become more
difficult to achieve. Calculating percent removals in these situations will produce deceiving
results and lead to unwarranted perceptions about the worth of the BMP. Influent and effluent
concentrations should always be reported in these cases. Rushton (2002) suggests comparing the
BMP effluent concentrations to applicable water quality standards. Still, even when the BMP is
part of a integrated system, monitoring of individual BMPs may be necessary to produce data for
modeling. Watershed modeling often requires efficiency monitoring data for each BMP
separately for the development of reliable statistical (black-box) models and/or the
verification/calibration of process-based models.
5.3.1 Integration of Structural BMPs
Structural BMPs can be used together in an integrated approach in two distinct ways. Individual
BMPs can be: (1) strung together in a directly connecting manner, i.e., the treatment train
approach; or (2) operated separately, but used collectively in the protection of the same receiving
water or watershed system.
5 - 43
Many possible treatment train options exist. Generally, smaller BMPs are used as pretreatment
for and/or conveyance to larger BMPs. Using smaller BMPs as pretreatment has the added
benefit of extending the maintenance intervals and useful life of the primary BMP. Examples of
such BMP treatment trains are as follows:
•
vegetated filter strips 6 sand filter or infiltration basin,
•
grassed swales 6 dry extended-detention pond, wet pond, or wetland,
•
bioretention 6 dry extended-detention pond or infiltration basins, and
•
manufactured products for stormwater inlets or catchbasins and catchbasin inserts
6 dry extended-detention ponds or infiltration basins.
Using structural BMPs in treatment trains, such as those listed above, can improve both the water
quality as well as total runoff volume of the final effluent, thereby minimizing the effects on
receiving waters (U.S. EPA, 1999a). Unlike the unit processes in conventional drinking water or
wastewater treatment trains that are engineered to perform a specific task, such as sedimentation
or disinfection, individual structural BMPs are often not limited to one pollutant removal
mechanism, as discussed in Chapter 2. However, the most effective combinations are made
between BMPs with different dominant pollutant removal mechanisms, such as integration of a
sand filter (filtration) with a wet retention pond (sedimentation). Combining strong water
quantity control BMPs (e.g., dry detention ponds) with strong water quality control BMPs (e.g.,
sand filters) will also make effective treatment trains.
The multi-chambered treatment train (MCTT) is an example of an engineered BMP that
integrates several pollutant removal mechanisms used by various BMPs (Greb et al., 2000). This
manufactured device is akin to a miniature wastewater treatment plant with three main chambers,
a grit chamber, a settling chamber, and a media filter chamber. The system approximates a three
component treatment train starting with a catchbasin, then a small dry extended-detention pond
and polishing effluent with a media filter. The MCTT is designed to be installed underground
and one study found appreciable pollutant removals of TSS, total phosphorus, and total
recoverable phosphorus (Greb et al., 2000). However, this treatment train will not provide
significant water quantity control.
The Staten Island Bluebelt program discussed in Section 5.2 is an example of separate, but
collectively used structural BMP integration. Their pocket wetlands are a combination of ponds
and shallow marsh lands (Vokral et al., 2001). These BMPs were engineered to act as a buffer
between piped runoff and natural streams or drainage corridors. Individually each BMP provides
some pollutant removal and peak runoff attenuation, but combined they provide a very effective
system for stream protection. Also, since the individual parcels of land are often adjacent to each
other, the Bluebelt BMP system provides a continuous corridor for a healthy wildlife habitat.
5.3.2 Integration of Nonstructural BMPs
Given the shear number of nonstructural BMP options, listing every possible combination of
nonstructural BMPs would be impractical. Instead, four examples of nonstructural BMP
integration are presented in the following paragraphs.
5 - 44
The benefit of the combined effect of sensitive construction planning and implementation of low
impact development (LID) techniques can be seen in a comparison study produced by Finnemore
(1982). The integrated set of BMPs used were open space requirements, cluster development of
residential areas, strict avoidance of construction on areas with slopes steeper than 25%, and
proper construction practices including, check dams and erosion baffles, lining drainage
channels, filter fences and berms, and restoration of native vegetation. Suspended sediment
loads were measured before and after residential development of two sites in Lake Tahoe,
California. The well-planned, water sensitive site increased sediment loading by a factor of two
and receiving water studies of macro-invertebrate communities detected a negligible impact,
whereas the poorly planned site which implemented no BMPs increased sediment loadings by a
factor of 107, causing an observed 34% decrease in macro-invertebrate density and a 54%
decrease in the number of macro-invertebrate species. Thus, it was estimated that the benefit of
this combination of construction site BMPs and LID measures was a 98% reduction in suspended
sediment loading. The author of the study did caution that both the sites chosen for their study
were extreme cases and that the impacts of most new development, whether BMPs are
implemented or not, likely will fall somewhere in between the two they presented.
A municipality in Oklahoma instituted an integrated stormwater management program including
the following BMPs; illegal discharge elimination program; litter collection campaign; illegal
dumping minimization program; hazardous waste collection program; storm drain stenciling
program; and an extensive advertizing campaign (Taylor and Wong, 2002a). Stormwater quality
was monitored for four years both before and after implementation of the program. Collectively
these BMPs produced average reductions in event mean concentrations of 13% TSS, 17% total
phosphorus, 18% total kjeldahl nitrogen, and at least 55% for the heavy metals, copper, lead, and
zinc. Although the data was not statistically validated, the results are quite promising.
When the effectiveness of education and outreach initiatives are viewed independently there is
little evidence that “information alone can significantly change actual behavior” (Taylor and
Wong, 2002a). However, use of an educational component as part of an integrated stormwater
management program is integral to the success of many nonstructural BMPs. For example,
ordinances will be ineffective unless preceded by an appropriate educational initiative (Taylor
and Wong, 2002a). People cannot be expected to comply with a law if they don’t know it exists.
A review of over 100 case studies found that the effectiveness of other nonstructural BMP
programs are often related to the effectiveness of education and outreach programs (Taylor and
Wong, 2002a). The success of most integrated programs should improve if a well executed
educational component is included. Yet, education and laws alone will not lead to compliance
unless followed up by a strong enforcement program for incentive. The bigger picture is that
ordinances by themselves are not a very effective BMP, but when integrated with education and
enforcement, their effectiveness greatly improves.
5.3.3 Integrating Structural and Nonstructural BMPs
In general, structural BMPs are used much more readily than nonstructural BMPs. Structural
BMPs are tangible entities that produce measurable results. They provide “the engineering
5 - 45
calculations necessary to demonstrate compliance with numerical ... criteria” (NVPDC, 1996).
However, most stormwater managers would agree that an effective stormwater management
program requires an optimum balance between both structural and nonstructural BMP types
(Taylor and Wong, 2002b). A comprehensive study of stormwater management programs
concluded that effective integration of many types of BMPs produce the most successful results.
Nonstructural practices that reduce the amount of runoff, such as cluster housing, reducing street
width, and alternative turn-about designs, can reduce size requirements or numbers of structural
BMPs. Nonstructural BMPs also can lessen pollutant loading thereby reducing the required
maintenance burden of structural BMPs.
For these reasons, the concept of LID has gained a considerable amount of support. LID is often
referred to as a single BMP, however, it is really an integrated concept containing both
distributed structural and nonstructural components. Typical structural BMPs incorporated into
LID include swales, vegetated filter strips, porous pavement, bioretention areas, and
disconnection of roof leaders into rain barrels, dry wells, or rain gardens. Nonstructural BMPs
commonly used in LID include construction site erosion and sediment controls, minimization of
impervious areas, and pollution prevention and good housekeeping practices. Integration of
these soft engineering distributed structural BMPs and nonstructural BMPs can result in a
considerable reduction in pollutant loading.
Zielinski (2000a) completed a modeling study to compare LID to two conventionally designed
residential developments with respect to percent imperviousness, stormwater runoff volume, and
nutrient export. In a low-density residential area, LID elements (called open space design in the
reference) included smaller lot size, narrower streets, shared driveways, a looped turnabout with
central bioretention cell instead of the conventional cul-de-sac bulb, natural area protections,
grassed swales, and improved septic tank design and installation. Modeling efforts showed
reductions in average annual total nitrogen and phosphorus loads of 46 and 50%, respectively,
and a 35% decrease in total impervious area over the conventional design without BMPs. In a
medium-density residential area the LID elements were similar, including smaller lot size,
narrower streets, vegetated buffer strips along waterways, shorter driveways, fewer sidewalks,
elimination of turnabouts by using a looped road design, grassed swales, and several bioretention
areas. A small central detention pond was also included in the open space design. This
combination of LID elements and the pond resulted in a 24% reduction in total impervious area
for the site. Furthermore, a significant reduction in average annual total nitrogen and phosphorus
loads of 45 and 60%, respectively, was estimated for the open space design, even though the
current conventional design included a large dry extended-detention pond.
In a related study Zielinski (2000b) explored LID practices for green parking lot designs. In one
scenario, the green parking lot design included the following elements: reduction in the number
and size of parking spaces, reduction in the width of the aisles, larger landscaping islands, some
planted with shade trees, others used as bioretention areas, grassed swales, porous pavement
(grid pavers), and a single wet retention pond to handle larger storms only. The conventionally
designed parking lot included two large wet ponds for stormwater treatment. Still, modeling
results estimated average annual total nitrogen and phosphorus load reductions of 45 and 47%,
5 - 46
respectively, and a 22% reduction in total impervious area over the conventional design without
BMPs. The results from the three scenarios investigated by Zielinski (2000a; 2000b), show that
through the integration of structural and nonstructural BMPs better protection of the watershed is
achieved than when each element is used individually.
LID supporters have highlighted other advantages of this micro-scale technique in addition to
water quality improvements. LID’s lot-level strategies can reduce hydrologic stresses of
development. Pre-development hydrology is replicated as closely as possible by managing the
stormwater runoff where it starts, sometimes eliminating the need for large scale centralized
BMPs such as retention ponds (Lai et al., 2003). Groundwater recharge is maximized by LID
practices (Clar et al., 2003). Finally, many of the techniques used in LID will reduce
development costs and associated maintenance costs, such as having sidewalks on only one side
of the street. Yet, LID is in its infancy compared to other BMPs and therefore it is not a proven
approach. Disadvantages of LID include conflicts with some areas’ building codes, difficulties
gaining public support for new, unproven techniques, difficulties in constructing and maintaining
and highly distributed structural controls, and the perception by some that LID promotes urban
sprawl (Clar et al., 2003).
REFERENCES
Atlanta Regional Commission (ARC). (2001). Georgia stormwater management manual,
Volume 2: technical handbook (first edition). Atlanta, GA. www.georgiastormwater.com.
August 2001.
Barraud, S., Y. Azzout, F.N. Cres, and B. Chocat. (1999). Selection aid of alternative techniques
in urban storm drainage - proposition of an expert system. Water Science and Technology
39(4): 241 - 248.
Barrett, M.E. (2004a). Performance comparison of structural stormwater BMPs. Water
Environment Research: awaiting publication.
Barrett, M.E. (2004b). Retention pond performance: examples from the international stormwater
BMP database. In: Proceedings of the World Water and Environmental Resources
Congress - Critical Transitions in Water and Environmental Resources Management,
June 27 - July 1, 2004, Salt Lake City, UT.
Clar, M., B.J. Barfield, and S. Yu. (2003). Considerations in the design of treatment best
management practices (BMPs) to improve water quality. EPA 600/R-03/103. U.S.
Environmental Protection Agency, Office of Research and Development, Cincinnati, OH.
www.epa.gov/ORD/NRMRL/Pubs/600r03103/600r03103.htm October 2003.
Commonwealth of Virginia (COV). (1999). Virginia stormwater management handbook, Volume
I (first edition). Department of Conservation and Recreation, Division of Soil and Water
Conservation, Richmond, VA. http://ww.dcr.state.va.us/sw/docs/swm/Volume_I.pdf.
Accessed on March 24, 2004.
Florida Department of Environmental Protection (FDEP). (2003). Nonpoint source management
- urban stormwater program. Nonpoint Source Management Program, Tallahassee, FL.
5 - 47
http://www.dep.state.fl.us/water/nonpoint/urban1.htm. April 2003. Accessed on January
21, 2004.
Federal Highway Administration (FHWA). (2000). Stormwater best management practices in an
ultra-urban setting: selection and monitoring. FHWA-EP-00-002. U.S. Department of
Transportation, Federal Highway Administration, Washington, DC.
http://www.fhwa.dot.gov/environment/ultraurb/index.htm.
Finnemore, E.J. (1982). Stormwater pollution control: best management practices. Journal of
Environmental Engineering Division, Proceedings of the American Society of Civil
Engineers 108(EE5): 706 - 721.
Frost, A. (2003). Can BMPs really work in Scotland?, http://www.sepa.org.uk/guidance/urbandrainage/bmps.htm. Scottish Environmental Protection Agency (SEPA), Scotland, UK.
GeoSyntec Consultants Urban Drainage and Flood Control District and Urban Water Resources
Council of ASCE (GeoSyntec and ASCE). (2002). Urban stormwater BMP performance
monitoring: a guidance manual for meeting the national stormwater BMP database
requirements. EPA 821/B-02/001. U.S. Environmental Protection Agency, Office of
Water, Washington, DC.
Greb, S.R., R.T. Bannerman, S.R. Corsi, and R.E. Pitt. (2000). Evaluation of the multichambered
treatment train, a retrofit water-quality management practice. Water Environment
Research 72(2): 207 - 216.
Lai, F.-h., C.-Y. Fan, L. Shoemaker, and R. Field. (2003). Development of a decision-support
framework for placement of BMPs in urban watersheds. In: National TMDL Science and
Policy 2003 Specialty Conference, Chicago, IL. Water Environment Federation.
Law, N.L. and L.E. Band. (1998). Performance of urban stormwater best management practices,
http://www.unc.edu/depts/geog/them/projects/BMP.html. University of Carolina - Chapel
Hill, Department of Geology, Chapel Hill, NC.
New Jersey Department of Agriculture, New Jersey Department of Community Affairs, New
Jersey Department of Environmental Protection, and New Jersey Department of
Transportation (NJDA et al.). (2000). DRAFT revised manual for New Jersey: best
management practices for control of nonpoint source pollution from stormwater (Fifth
draft). http://www.state.nj.us/dep/watershedmgt/bmpmanual.htm. May 3, 2000.
Commission, N.V.P.D. (NVPDC). (1996). Nonstructural urban BMP handbook - a guide to
nonpoint source pollution prevention and control through nonstructural measures.
Prepared for the Department of Conservation and Recreation, Division of Soil and Water
Conservation, Annandale, Virginia. http://novaregion.org/NSBMP1.
New York State Department of Environmental Conservation (NYSDEC). (2001). New York state
stormwater management design manual (Draft). Division of Water, Albany, NY.
http://www.dec.state.ny.us/website/dow/swmanual. October 2001.
Parikh, P. (2003). A history and overview of the clean water act and its application to stormwater
runoff. National Risk Management Research Laboratory Seminar Series, November 6,
2003, Cincinnati, OH.
Pitt, R., M. Lalor, R. Field, D.D. Adrian, and D. Barbe. (1993). Investigation of inappropriate
pollutant entries into storm drainage systems. EPA 600/R-92/238. U.S. Environmental
Protection Agency, Office of Research and Development, Cincinnati, OH. January 1993.
5 - 48
Roesner, L.A. (1999). Urban runoff pollution summary thoughts - the state-of-practice today and
for the 21st century. Water Science and Technology 39(12): 353 - 360.
Rushton, B. (1998). Processes that affect stormwater pollution, http://www.stormwaterresources.com/library/001PPocesses.PDF. Stormwater News, Cocoa Beach, Florida.
Rushton, B. (2002). BMP monitoring: methods and evaluations, http://www.stormwaterresources.com/library.htm#BMPs (#147PL). Stormwater News, Cocoa Beach, Florida.
Schueler, T.R. (1987). Controlling urban runoff: a practical manual for planning and designing
urban BMPs. Metropolitan Washington Council of Governments, Department of
Environmental Programs, Washington, DC. July, 1987.
Selvakumar, A. and M. Borst. (2003). Effects of land use and season on microorganism
concentrations in urban stormwater runoff. In: National TMDL Science and Policy 2003
Specialty Conference, Chicago, IL. Water Environment Federation.
Sieker, H. and M. Klein. (1998). Best management practices for stormwater-runoff with
alternative methods in a large urban catchment in Berlin, Germany. Water Science and
Technology 38(10): 91 - 97.
Strecker, E., P. Mangarella, and N. Brandt. (2002). A natural treatment system master plan
approach to meeting TMDLs: case study of San Diego Creek, Orange County, California.
In: National TMDL Science and Policy Specialty Conference, November 13 - 16, 2002,
Phoenix, AZ. Water Environment Federation.
Strecker, E.W., M.M. Quigley, B. Urbonas, and J. Jones. (2004). Analyses of the expanded
EPA/ASCE international BMP database and potential implications for BMP design. In:
Proceedings of the World Water and Environmental Resources Congress - Critical
Transitions in Water and Environmental Resources Management, June 27 - July 1, 2004,
Salt Lake City, UT.
Stribling, J.B., E.W. Leppo, J.D. Cummins, J. Galli, S. Meigs, L. Coffman, and M.-S. Cheng.
(2001). Relating instream biological condition to BMPs in watershed. In: Linking
Stormwater BMP Designs and Performance to Receiving Water Impact Mitigation,
August 19 - 24, Snowmass Village, CO. American Society of Civil Engineers.
Sullivan, D., R. Field, and C.-Y. Fan. (2003). The complications of monitoring treatment BMPs.
In: National TMDL Science and Policy 2003 Specialty Conference, Chicago, IL. Water
Environment Federation.
Taylor, A. and T. Wong. (2002a). Non-structural stormwater quality best management practices
- a literature review of their use, value, cost, and evaluation. EPA Victoria Technical
Report 02/11. Cooperative Research Center for Catchment Hydrology, Victoria,
Australia. http://www.catchment.crc.org.au/pdfs/Technical200213.pdf December 2002.
Taylor, A. and T. Wong. (2002b). Non-structural stormwater quality best management practices
- an overview of their use, value, cost, and evaluation. EPA Victoria Technical Report
02/13. Cooperative Research Centre for Catchment Hydrology, Victoria, Australia.
http://www.catchment.crc.org.au/pdfs/technical200211.pdf December 2002.
United States Environmental Protection Agency (U.S. EPA). (1999a). Preliminary data summary
of urban storm water best management practices. EPA 821/R-99/012. Office of Water,
Washington, DC. August 1999.
United States Environmental Protection Agency (U.S. EPA). (1999b). Stormwater management
fact sheet - non-storm water discharges to storm sewers. EPA 832/F-99/022. Office of
5 - 49
Water, Washington, DC. http://www.epa.gov/owm/mtb/nonstorm.pdf. September 1999.
Accessed on January 26, 2004.
United States Environmental Protection Agency (U.S. EPA). (2000). Storm water Phase II final
rule: an overview. EPA 833/F-00/001. Office of Water, Washington, DC. January 2000.
United States Environmental Protection Agency (U.S. EPA). (2002a). National management
measures to control nonpoint source pollution from urban areas (Draft). EPA 842/B02/003. Office of Water, Office of Wetlands, Oceans, and Watersheds, Nonpoint Source
Control Branch, Washington, DC. September 2002.
United States Environmental Protection Agency (U.S. EPA). (2002b). Illicit discharge detection
and elimination - wastewater connections to the storm drain system. Office of Water,
Office of Wastewater Management, Washington, DC.
http://cfpub.epa.gov/npdes/stormwater/menuofbmps/illi_7.cfm. August 2002. Accessed
on January 26, 2004.
United States Environmental Protection Agency (U.S. EPA). (2003a). Post-construction storm
water management in new development and redevelopment - dry extended detention
pond. Office of Water, Washington, DC.
http://cfpub.epa.gov/npdes/stormwater.menuofbmps/post_26.cfm. Accessed on August
27, 2003.
United States Environmental Protection Agency (U.S. EPA). (2003b). National menu of best
management practices for storm water Phase II. Office of Water, Washington, DC.
http://cfpub.epa.gov/npdes/stormwater/menuofbmps/menu.cfm. Accessed on August 27,
2003.
United States Environmental Protection Agency (U.S. EPA). (2003c). Phase of the NPDES storm
water program. http://cfpub.epa.gov/npdes/stormwater/swphases.cfm. Accessed on
March 30, 2004.
Urbonas, B. (2000). Assessment of stormwater best management practice effectiveness (chapter
7). In: Innovative urban wet-weather flow management systems, Field, R., J.P. Heaney,
and R. Pitt, (editors). Technomic Publishing Co., Inc., Lancaster, PA. Pages 255 - 300.
Vokral, J., D. Gumb, R.D. Smith, and S. Mehrotra. (2001). Staten Island bluebelt program - a
natural solution to environmental problems. Stormwater 2(3): 58 - 69.
Wayland, R.H. and J.A. Hanlon. (2002). Memorandum: establishing total maximum daily load
(TMDL) wasteload allocations (WLAs) for storm water sources and NPDES permit
requirements based on those WLAs, http://www.epa.gov/npdes/pubs/final-wwtmdl.pdf.
U.S. Environmental Protection Agency, Office of Water, Washington, DC. Accessed on
August 18, 2004.
Water Environment Federation and American Society of Civil Engineers (WEF and ASCE).
(1998). Urban runoff quality management. 1-57278-039-8 and 0-7844-0174-8. Water
Environment Federation, Alexandria, VA.
Zielinski, J. (2000a). The benefits of better site design in residential subdivisions. In: The
Practice of Watershed Protection, Schueler, T.R. and H.K. Holland, (editors). The Center
for Watershed Protection, Ellicott City, MD. Pages 263 - 275.
Zielinski, J. (2000b). The benefits of better site design in commercial development. In: The
Practice of Watershed Protection, Schueler, T.R. and H.K. Holland, (editors). The Center
for Watershed Protection, Ellicott City, MD. Pages 277 - 286.
5 - 50
6
BMP Costs
Ari Selvakumar
CONTENTS
6.0
6.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-1
BMP Cost Estimating Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-2
6.1.1 Bottom-Up Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-3
6.1.2 Analogy Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-3
6.1.3 Expert Opinion Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-3
6.1.4 Parametric Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-3
6.2
BMP Cost Analyses Elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-3
6.2.1 Total Costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-3
6.2.1.1 Capital Costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-4
6.2.1.2 Design, Permitting, and Contingency Costs . . . . . . . . . . . . . . . . . . . . 6-5
6.2.1.3 Operation and Maintenance (O&M) Costs . . . . . . . . . . . . . . . . . . . . . . 6-6
6.2.1.4 Land Costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-7
6.2.1.5 Inflation and Regional Cost Adjustments . . . . . . . . . . . . . . . . . . . . . . 6-8
6.2.2 Life Cycle Costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-11
6.3
Economic Benefits of BMPs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-12
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-15
6.0
INTRODUCTION
Cost is an integral part of BMP design. The purpose of this chapter is to provide information on
how to estimate the cost of structural BMPs. Costs associated with nonstructural BMPs are not
included here as they are generally not as easily quantified as structural BMPs due to their
indirect and highly variable implementation levels. The cost for constructing a structural BMP
depends on many factors. They include the time of year, site conditions and topography,
accessibility of equipment, economics of scale, and government regulations. Out of these
factors, site conditions and topography are usually the most influential. Site preparation costs
may be greatly reduced if existing conditions and vegetation are carefully integrated into the
design of any BMP; e.g., a natural depression could be developed into some type of detention
pond (Ferguson et al., 1997).
It is normally less expensive from a unit cost standpoint to construct a larger project than a
6-1
smaller one. This is because large portions of the costs are involved in mobilizing and
demobilizing a project. Both overestimating and underestimating can occur by not considering
the magnitude of the project (Ferguson et al., 1997). However, a larger project usually involves
significantly more land costs than does distributing BMPs into already set aside landscaping
areas.
The cost of constructing any BMP is variable and can be substantial. Several documents have
been published that address cost estimating for BMPs, but most of these report only construction
costs (Young et al.,1996; Sample et al., 2003). In addition, costs are often documented as base
costs and do not include land costs, which according to the U.S. EPA (1999) is the largest
variable influencing overall BMP cost.
Design, operation, and maintenance (O&M) are other significant cost considerations discussed in
only a few sources. Ferguson et al. (1997) published a fairly comprehensive BMP cost
estimating guide as part of the Rouge River restoration project that includes construction and
O&M costs for many BMPs. This guide could be used outside the Rouge River watershed
provided adjustments are made for local labor rates and associated real estate costs. Young et al.
(1996) compiled the results of past highway runoff research into a single volume users manual
for highway practitioners which provide construction cost data for each BMP, but no cost data
for design and maintenance. The Southeastern Wisconsin Regional Planning Commission
(SWRPC) (1991) documents the most comprehensive analysis of construction and maintenance
costs. The Water Environment Research Foundation (WERF) has summarized construction and
maintenance costs for commonly used BMPs (WERF, 2003). The wide range of cost data
reported in the literature indicates that much more information is needed in this area. Identifying
the primary element that drives construction cost upward is a high priority and could help reduce
the initial cost associated with BMPs.
In general, initial construction cost and maintenance cost were the factors considered in
computing the cost effectiveness of BMPs. Recent studies have suggested that using
construction and maintenance costs to compute cost effectiveness is insufficient. Landphair et
al. of Texas Transportation Institute (2000) proposed that a real measure of cost effectiveness
should include design, maintenance, and construction costs as well as pollutant removal
performance of a selected BMP. Their research suggested that a cost index of this kind was only
meaningful if special site considerations and land costs were ignored.
6.1
BMP COST ESTIMATING METHODS
Four common methods of cost estimation are as follows (DOD, 1995):
•
•
•
•
bottom-up method,
analogy method,
expert opinion method, and
parametric method.
6-2
6.1.1 Bottom-Up Method
The bottom-up approach estimates costs on an item-by-item basis. Detailed methods typically
rely on quantity take-offs and compiled sources of unit cost data for each item, taken from either
a built-in database or other sources (e.g., cost-estimating references). This method is used when
design information is available.
6.1.2 Analogy Method
This technique uses the cost of previously completed projects, as a comparison to the cost of the
proposed project, to arrive at a final cost estimate. The actual costs from the completed project
are extrapolated to estimate the cost of the proposed project (DOD, 1995). This method provides
estimates by using the system or component level, which are obtained from using actual project
costs from past experiences. The disadvantage of this method occurs when projects similar to
the one in question do not exist, or the accuracy of the available data is questionable, or
extrapolating the entire project cost does not accurately reflect evenly scaled costs.
6.1.3 Expert Opinion Method
Experts in this field can be consulted to provide a cost estimate for the project based upon their
past experience and understanding of the proposed undertaking. The advantage of this is that the
experts have the knowledge to account for the differences between past project experiences and
the proposed project requirements, pointing out areas contributing to the cost estimate that may
be overlooked. They can also factor in the impacts produced by new technologies or
applications. The disadvantage of this technique is that the estimate is confined by the
judgement and expertise of the consulted experts (DOD, 1995).
6.1.4 Parametric Method
The parametric approach relies on relationships between cost and design parameters. These
relationships are usually statistically-based or model-based. Statistically-based approaches rely
on “scaled-up” or “scaled-down” versions of projects where historical cost data is available.
Model-based approaches utilize a generic design that is linked to a cost database and adjusted by
the user for site-specific information. This method, also known as top-down estimating, is used
when design information is not available (U.S. EPA, 2000).
6.2
BMP COST ANALYSES ELEMENTS
6-3
6.2.1 Total Costs
The total costs of BMPs include both capital (construction and land) and annual O&M costs.
The first category of costs only occur in the year when the BMP is installed unless retrofits or
up-sizing occur. The other costs may occur either yearly (e.g., operation cost) or some other
frequency throughout the life of the BMP to maintain it.
6.2.1.1 Capital Costs
Capital costs are those expenditures that are required to construct a BMP. They are exclusive of
costs required to operate or maintain the BMP throughout its lifetime. Capital costs consist
primarily of expenditures initially incurred to build or install the BMP (e.g., land cost,
construction of a wetland, and related site work).
Capital costs include all land, labor, equipment and material costs, excavation and grading,
control structure, sediment control, landscaping, and appurtenances. Capital costs also include
expenditures for professional/technical services that are necessary to support the construction of
a BMP. The cost of constructing any BMP is variable and depends largely on site conditions and
drainage area. For example, if a BMP is constructed in very rocky soils, the increased
excavation costs may substantially increase the cost of construction. Land costs, which can be
the largest variable influencing overall BMP cost, vary greatly from site to site.
Capital costs typically can be estimated using equations based on the size or volume of water to
be treated (U.S. EPA, 1999). Stormwater practices have certain spatial requirements for surface
area or volume necessary to treat a quantity of stormwater from a watershed and this variation is
incorporated into the equations (U.S. EPA, 1999). These equations typically follow the single
determinant equation as shown below. This equation is linear in the log transform.
C = aP b
Where C is the estimated cost ($); P is the determinant variable (area or volume); and a and b are
statistical coefficients determined from regression analysis. The exponent b represents the
economics of scale factor. If b < 1, the unit cost decreases as size increases. When b = 1, the
equation simplifies to a linear relationship and no economics-of-scale are present. If b < 1, then
diseconomics of scale exist. Table 6-1 shows the unit costs of commonly used BMPs as reported
in the literature. Unless indicated otherwise, all costs are expressed in terms of 2002 dollars
based on the Engineering News Record (ENR) average annual cost index of 6538 (ENR, 2002).
The cost equations show that the capital costs of most BMPs correlate well with basin volume.
The same was noted by Koustas and Selvakumar (2003).
Traditionally, capital costs were calculated using the standard estimation guides such as the R.S.
Means Building Construction Cost Data Handbook used by many engineers. The Means book
6-4
provides unit cost data for materials of construction, labor, equipment, installation, and
excavation for cities across the U. S. However, linear equations are gaining popularity as they
offer a way to replace a cost database with a single equation. These equations are based on data
gathered from existing projects and assumptions that similar installations should cost about the
same.
Table 6-1. Base Capital Costs (Without Land Costs) for Commonly Used BMPs
BMP Type
Base Capital Costs ($)
Reference
Detention Ponds/Dry
Extended Detention Ponds
C = 60,742V0.69; V in Mgal
Young et al., 1996
C = 12.4V0.76; V in ft3
Brown and Schueler, 1997
Wet Ponds/Retention Basins
C = 67,368V0.75; V in Mgal
Young et al., 1996
C = 24.5V0.71; V in ft3
Brown and Schueler, 1997
Constructed Wetlands
C = 30.6V0.71; V in ft3
U.S. EPA, 2003
Infiltration Trenches/Filter
Drains/Soakaways
C = 173V0.63; V in ft3
Young et al., 1996
C = 5V; V in ft3
Brown and Schueler, 1997
Infiltration Basins
C = 16.9V0.69; V in ft3
Young et al., 1996
Sand and Organic Filters
C = KA; A in acres; K ranges
from 12,369 to 24,738
Young et al., 1996
Vegetated Swales
$0.25 to $0.50/ft2
WERF, 2003
Vegetated Buffer Strips
$0.30 to $0.70/ft2
WERF, 2003
Porous Pavement
$2 to $3/ft2
U.S. EPA, 2003
Bioretention
$3 to $4/ft2
Coffman, 1999
C = 7.3V0.99; V in ft3
U.S. EPA, 2003; Brown and
Schueler, 1997
$8,000 to $24,000
Young et al., 1996
$2,000 to $3,000/basin for
precast basins
U.S. EPA, 2003
$400 to $10,000/basin for
drop-in retrofits
U.S. EPA, 2003
Water Quality Inlets
(enhanced catch basins)
Note:
Costs in December 2002 dollars. Cost of land acquisition not included.
V = BMP Volume and A = BMP Area.
6.2.1.2 Design, Permitting, and Contingency Costs
6-5
Design and permitting costs include costs for site investigations, surveys, design and planning of
a BMP (Novotny and Chesters, 1981). Contingency costs are simply any unexpected costs
incurred during the development and construction of a BMP. These costs are usually expressed
as a fraction of the base capital costs and have been considered as a uniform percentage for
BMPs (U.S. EPA, 1999). Design, permitting, and contingency costs are reported in Table 6-2.
These costs are generally only estimates, based on the experience of designers. Design,
permitting and contingency costs can be a significant portion of the total capital cost of a BMP
construction project particularly in areas where endangered species occur in contentious projects.
In addition, cost of erosion and sediment control represents about 5 to 7% of the base capital
costs (Brown and Schueler, 1997; U.S. EPA, 1999).
Table 6-2. Design, Permitting, and Contingency Costs
Additional Percentage
(fraction of base capital costs)
Notes
Source
25 %
Includes design, permitting
and contingency fees
Wiegand, et al. 1986; CWP,
1998; and U.S. EPA, 1999.
32 %
Includes design, permitting,
contingency fees and
erosion and sediment
control costs
Brown and Schueler, 1997;
U.S. EPA, 1999; and CWP,
1998.
6.2.1.3 Operation and Maintenance (O&M) Costs
O&M costs are those post-construction costs necessary to ensure or verify the continued
effectiveness of a BMP. Like engineering design and construction, proper O&M is important in
order to ensure that BMPs will function properly. These costs are estimated mostly on an annual
basis.
Annual O&M costs include labor, materials, energy, and equipment required for proper
operation and functionality of a BMP facility. These include costs for landscape maintenance,
structural maintenance, infiltration maintenance, sediment removal, and basin debris and litter
removal.
O&M costs are divided into two main categories: aesthetic and functional (U.S. EPA, 1999).
Functional maintenance is important for performance and safety reasons, while aesthetic
maintenance is important primarily for public acceptance of BMPs. Aesthetic appearance is
more important for BMPs that are visible.
O&M costs are calculated on an annual basis throughout the life of a BMP facility. These costs
are seldom available on a comprehensive basis and have been expressed as a fraction of the base
6-6
capital costs. O&M costs, expressed as annual percentages of capital costs, as reported in
literature, are presented in Table 6-3. O&M costs to maintain a stormwater BMP over a 20 to
25-year period are roughly equal to its initial construction cost (Weigand et al., 1986). Few
property owners and homeowner associations are fully aware of the need for and magnitude of
stormwater maintenance costs, and most fail to regularly perform routine and non-routine
maintenance tasks. It is likely that performance and longevity of many BMPs will decline
without adequate maintenance. The information on the maintenance of BMPs is limited, even
though the long term maintenance costs can be an important component of the life cycle cost of
the facility.
Table 6-3. Annual Operation & Maintenance Costs
BMP Type
Annual Maintenance Costs
(% of Construction Cost
without Land Cost)
Detention Ponds/Dry
<1%
Extended Detention
Ponds
Wet Ponds/Retention
3 to 6%
Basins
3 to 6%
Constructed Wetlands
2%
Infiltration Trench
Infiltration Basin
5 to 20%
1 to 3%
5 to 10%
Sand Filters
Grassed Swales
Vegetated Buffer
Strips
11 to 13%
5 to 7%
$350/acre/year
Source
Wiegand et al., 1986; Schueler,
1987; SWRPC, 1991
Brown and Schueler, 1997;
SWRPC, 1991
Wiegand et al. 1986; Schueler,
1987; SWRPC, 1991
Livingston et al. 1997; Brown
and Schueler, 1997
Schueler, 1987; SWRPC, 1991
Livingston et al. 1997; SWRPC,
1991
Wiegand et al. 1986; Schueler,
1987; SWRPC, 1991
Livingston et al. 1997; Brown
and Schueler, 1997
SWRPC, 1991
SWRPC, 1991
6.2.1.4 Land Costs
Land cost is of critical importance because it constitutes a significant, if not major, component of
total costs. The cost of land is site specific and extremely variable both regionally and by
surrounding land use. For example, many suburban jurisdictions require open space allocations
within the developed site, reducing the effective cost of land for BMPs to zero (Schueler, 1987).
On the other hand, the cost of land may far outweigh construction and design costs in ultra urban
settings. It should be noted that in some cases where a BMP can be incorporated into a planned
6-7
or required landscaped area, there is no additional cost. For example, Portland, Oregon has a
15% landscaping requirement for multi-family, commercial, and light-industrial developments.
Automatically, there is enough land for stormwater BMPs. The land requirement varies
considerably depending on the type of BMP as shown in the Table 6-4 below.
Table 6-4. Relative Land Consumption of Stormwater BMPs
BMP Type
Land Consumption
(% of Impervious Area of the
Watershed)
Retention Basin
2 to 3%
Constructed Wetland
3 to 5%
Infiltration Trench
2 to 3%
Infiltration Basin
2 to 3%
Porous Pavement
0%
Sand Filters
0 to 3%
Bioretention
5%
Swales
Filter Strips
10 to 20%
100%
Source: U.S. EPA, 1999
6.2.1.5 Inflation and Regional Cost Adjustments
Any costs reported in the literature need to be adjusted for inflation and regional differences. To
adjust for inflation, the ENR construction cost index history data can be used to adjust BMP cost
data to current year dollars. The ENR index is published every month. The ENR index for the
years 1990 to 2003 is shown in Table 6-5. To obtain current cost, the total capital BMP cost data
is multiplied by current year index and divided by base year indexes. The total cost increase
resulting from inflation can be calculated using the following equation:
Capital Cost (adjusted) =
Capital cost (base year) x ENR cost index for the current year
ENR cost index for the base year
6-8
Table 6-5. ENR Annual Average Construction Cost Index History
Year
Construction Cost Index
1990
4732
1991
4835
1992
4985
1993
5210
1994
5408
1995
5471
1996
5620
1997
5826
1998
5920
1999
6059
2000
6221
2001
6343
2002
6538
2003
6695
Source: Engineering News Record (ENR, 2003)
Total capital costs for the construction of a BMP facility vary based on the region of the county
or state in which the facility is located. Regional cost differences reflect differences in the cost
of living, labor, and material costs for a particular location (U.S. EPA, 1999). Regional cost
variations are perhaps the most difficult costs to estimate. Cost data can be adjusted for regional
cost variation by using the ENR twenty city construction cost indexes for December 2002 (ENR,
2002). The twenty city construction cost index for December 2002 is shown in Table 6-6. The
cost data adjusted for inflation is multiplied by the twenty city index factor (city index/national
index for December 2002) of the city closest to the region of study to adjust for the regional
variation.
Since the amount of regional rainfall may impact costs, further adjustment to cost is necessary.
Based on a methodology presented by the American Public Works Association (APWA, 1992), a
cost adjustment factor is assigned to each EPA rainfall region (shown in Figure 6-1). A regional
cost adjustment factor based on regional rainfall zones is shown in Table 6-7. The cost adjusted
for inflation and regional variation is reconciled for the rainfall bias by dividing the cost by the
6-9
rainfall factor. For example, data for Florida (Region 3) is divided by 0.67 to adjust for the
rainfall bias.
Table 6-6. Twenty City Construction Cost Index Regional Factors for December 2002
City
Index December 2002
Regional Factor
Atlanta
4189
0.64
Baltimore
4580
0.70
Birmingham
4686
0.71
Boston
7546
1.15
Chicago
7965
1.21
Cincinnati
6156
0.94
Cleveland
7067
1.08
Dallas
3895
0.59
Denver
4744
0.72
Detroit
7654
1.17
Kansas City
6782
1.03
Los Angeles
7403
1.13
Minneapolis
7621
1.16
New Orleans
3906
0.60
New York
10009
1.52
Philadelphia
8226
1.25
Pittsburgh
6419
0.98
St. Louis
7197
1.10
San Francisco
7644
1.16
Seattle
7562
1.15
Cost regional and inflation adjusted (December 2002) = Cost inflation adjusted (December 2002) x Regional Factor
Source: Engineering News Record (ENR, 2003); National Index for December 2002 is 6563
6-10
Table 6-7. Precipitation Regional Cost Adjustment Factors
Rainfall Zone
1
2
3
4
5
6
Adjustment
1.12
0.90
0.67 0.92 0.67 1.24
Factor
7
8
9
1.04
1.04
0.76
Source: U.S. EPA, 1999
Note: This is very approximate as there are large differences in precipitation patterns.
6.2.2 Life Cycle Costs
Life cycle costs are costs incurred from “cradle to grave”. Life cycle cost refers to the total
project cost across the life span of a BMP, including design, construction, O&M, and closeout
activities. It includes the initial capital costs and the present worth of annual O&M costs that are
incurred over time, less the present worth of the salvage value at the end of the service life
(Sample et al., 2003). Life cycle cost analysis can be used to choose the most cost effective
BMP from a series of alternatives so that the least long term cost is achieved.
A present worth analysis of future payment involves four basic steps:
1.
2.
3.
4.
define the period of analysis,
calculate the annual expenditures or cash outflow for each year of the project,
select a discount rate to use in the present worth calculation, and
calculate the present worth.
The present worth of a future payment is calculated using the following equation:
PW =
xt
(1 + i) t
where xt is the payment in year t and i is the discount rate. The denominator, (1 + i)t, is referred
as a “discount factor.” For a series of future payments, the total present worth from 1 to n years
would be calculated as:
PWtotal =
t=n
∑ (1 + i)
t =1
6-11
xt
t
t=n
where,
∑ (1 + i)
t =1
1
t
is referred as the multi-year discount factor. When the annual cost, xt, is
constant over a period of years, the calculation of present worth can be simplified by using a
multi-year discount factor.
Calculation Example - Wet Pond
In order to present how to calculate the life cycle cost of a BMP, a wet pond with a 1 acre-foot
facility is selected. From Table 6-1, the construction cost for a wet pond can be calculated using
the equation:
C = 24.5V 0.71
For a 1 acre-foot facility, the construction cost is equal to $45,700. From Table 6-2, the cost for
design, permitting, and contingency is 25% of the construction cost, which is equal to $11,425.
For wet ponds, the annual cost of routine maintenance is typically estimated at about 3 to 6% of
the construction cost (Table 6-3). Assuming 5%, the annual cost of maintenance becomes
$2,285. Ponds are long-lived facilities (typically longer than 20 years). Assuming a period of
analysis of 20 years and a discount rate of 7%, the multi-year discount factor is equal to 10.594.
The present worth of maintenance for 20 years is calculated as 10.594 x $2,285 or $24,207.
Thus, for this example the life cycle cost equals the construction cost ($45,700), plus the design,
permitting, and contingency cost ($11,425) and the present worth maintenance cost ($24,207)
totaling $81,332. Land cost, which is extremely variable is not included.
6.3
ECONOMIC BENEFITS OF BMPS
BMPs can present several tangible economic benefits in spite of their construction, O&M costs.
BMPs have the ability to mitigate potential downstream flooding associated with medium to
larger storms. BMPs reduce flooding downstream from developments by reducing the peak
flows and stretching out the flow over a long time. Avoiding floods eliminate property damages.
Benefits of avoiding flood damages are relatively easy to estimate. Reducing peak flows can
also potentially reduce capital and maintenance costs of drainage infrastructure.
BMPs can protect water resources downstream from urban development, which can prevent
impact on human health through direct contact from swimming or through contamination of
seafood. The economic benefits of avoiding human health problems include swimming and
recreational benefits as well as saved medical costs. One study in Saginaw, Michigan estimated
6-12
that the swimming and beach recreation benefits associated with a CSO retention project
exceeded seven million dollars (U.S. EPA, 1998). BMPs also reduce the cost of water quality
improvement.
BMPs are beneficial to developers as they create a “water” feature effect and people have a
strong emotional attachment to water, arising from its aesthetic qualities– tranquility, coolness,
and beauty. BMPs within developments can be used as marketing tools to set the tone for the
entire project. BMPs that are visually aesthetic and safe for children can lead to increased
property values. As such, beautification of land areas adjacent to BMPs should be considered an
integral part of planning by developers. It was found that developers could charge per lot a
premium of up to $10,000 for homes situated next to well-designed stormwater ponds and
wetlands (U.S. EPA, 1995). A study conducted by the National Association of Home Builders
indicates that the proximity to water raises the value of a home by up to 28% (U.S. EPA, 1995).
If the BMPs are developed to allow passive recreation (e.g., a walking path around a lake or
pond), the recreational area and the BMP can become the feature attraction when advertising the
property. BMPs with standing water often appear to be natural systems. Many developers have
capitalized on urban runoff regulations by designing aesthetic wet ponds and stormwater
wetlands and marketing them as if they were natural lakes, ponds, or wetlands. In addition, U.S.
EPA found that office parks and apartments next to a well-designed BMP could be leased or
rented at a considerable premium and often at a much faster rate. BMPs may require more than
periodic maintenance to preserve environmental and monetary benefits of “waterfront” lots.
However, the benefits of higher resale value and quality of life typically outweigh the combined
costs of the initial lot premium and annual maintenance fees charged by homeowners’
associations (U.S. EPA, 1995).
Stormwater BMPs such as grassed swales and bioretention areas are actually less expensive to
construct than enclosed storm drain systems and they also provide better environmental results
(NJDEP, 2000). Also, the more natural drainage systems eliminate the need for costly manholes,
pipes, trenches, and catch basins, while removing pollutants at the same time.
Stormwater BMPs must be maintained, and the cost burden falls on landowners or local
government. Over a 20 to 25 year period, the full cost to maintain a stormwater BMP is roughly
equal to its initial construction cost (Wiegand et al., 1986). Poorly maintained BMPs may be
unsightly due to excess algal growth or public littering. Few property owners and home owners’
associations are fully aware of the magnitude of stormwater BMP maintenance costs, and fail to
perform maintenance tasks.
Abatement or prevention of pollution from storm-generated flow is one of the most challenging
areas in the environmental engineering field. The facts of life - from an engineering standpoint
are difficult to face in terms of design and cost. Operational problems can be just as foreboding.
6-13
The full impacts of “marginal” pollution, particularly that caused by uncontrolled overflows,
must be recognized now and planning initiated to improve drainage system efficiencies and bring
all wastewater flows under control. Municipal programs with this objective cannot begin too
soon because corrective action is time-consuming. Efforts devoted to improved sewerage and
drainage systems will pay significant dividends in complete control of metropolitan wastewater
problems and pollution abatement. Research and development are making available important
answers on the most efficient and least costly methods needed to restore and maintain water
resources for maximum usefulness to man.
It is clear that abatement requirements for storm-flow pollution are here. Already, federal and
local governments have promulgated wet-weather flow treatment and control standards. Newly
developed and developing watershed districts can take a crucial opportunity and assess what has
transpired and determine their own best water management strategy.
To exemplify this, one can consider the water pollution control efforts in the U. S. Historically,
this nation has always approached water pollution control in a series and segmented manner with
respect to time and pollutant sources, respectively. The result is that the problem is still being
fought after more than 60 years of effort and billions of dollars of expenditures. Initially, they
abated sanitary sewage; first with primary treatment, and later only after a long time, with
secondary treatment. Somewhere in between attempts to control sanitary sewage, industrial
wastewater control became a requirement; however, pretreated industrial wastewaters are still
released during an overflow event. Only recently were we forced to control combined sewer
overflow, and now we are faced with requirements to abate separate-stormwater pollution. The
aforementioned historical approach to water pollution control has taken a very long time, and
only after trial and error of each individualized and fragmented approach was it learned that
receiving water pollution problems remain. If instead, an entire watershed or multi-drainage
area analysis was conducted earlier, a determination could have been made of the overall
pollution problem in the receiving water bodies, the pollution sources (or culprits) contributing
to the problem, and an optimized, integrated, area-wide program to correct the problem.
After the macro- (or large-scale watershed) analysis is conducted, an optimized determination of
what sources to be abated (or where to spend the monies) will be made. Then, with the resulting
information, a micro- (or drainage area/pollutant source and control) analysis can be performed.
There is one other important consideration that must be made, i.e., the reuse and reclamation of
stormwater for such beneficial purposes as aesthetic and recreational ponds, groundwater
recharge, irrigation, fire protection, and industrial water supply.
An optimal approach to integrated stormwater management is a total watershed or basin-wide
analyses including a macro- or large-basin-scale evaluation interfaced with a descretized microor small-catchment-scale evaluation involving the integration of: (1) all catchments or drainage
areas, tributaries, surrounding water bodies, and groundwater; (2) all pollutant source areas, land
6-14
uses, and flows; and (3) added storm-flow sludge and residual solids handling and disposal.
Flood and erosion control along with reuse and reclamation technology must also be integrated
with pollution control, so that the retention and drainage facilities required for flood and erosion
control can be simultaneously designed or retrofitted for pollution control and stormwater
reclamation.
In conclusion, knowledge of interconnecting basinwide waters and pollutant loads affecting the
receiving water body and the subsurface and groundwater will result in knowing how to get the
optimum water resource and pollution abatement and a much more expedient and cost-effective
water management program.
REFERENCES
American Public Works Association (APWA). 1992. A Study of Nationwide Costs to
Implement Municipal Storm Water Best Management Practices. Southern California
Chapter. Water Resource Committee.
Brown, W. and T. Schueler. 1997. The Economics of Storm Water BMPs in the Mid-Atlantic
Region. Center for Watershed Protection. Elliot City, MD, 41 pp.
Center for Watershed Protection (CWP). 1998. Cost and Benefit of Stormwater BMPs. Final
Report 9/14/98. EPA Contract 68-C6-0001 WA2-15, Task 6 Parson Engineering
Science.
Coffman, L., M. Clar, and N. Weinstein. 1998. Overview: Low Impact Development for
Stormwater Management. Unpublished.
ENR. 2002. Construction Economics. pp. 23. December 9.
ENR. 2003. Construction Cost Index History.
http://www.enr.com/features/conEco/costIndexes/constIndexHist.asp
Ferguson, T., R. Gignac, M. Stoffan, A. Ibrahim, and J. Aldrich. 1997. Rouge River National
Wet Weather Demonstration Project: Cost Estimating Guidelines, Best Management
Practices and Engineered Controls. Wayne County, MI.
Koustas, R.N. and A. Selvakumar. 2003. BMP Cost Analysis for Source Water Protection.
American Water Resources Association (AWRA) 2003 International Conference in New
York in June.
Landphair, H.C., J.A. McFalls, and D. Thompson. 2000. Design Methods, Selection, and CostEffectiveness of Stormwater Quality Structures. Report No. FHWA/TX-01/1837-1.
Texas Transportation Institute, The Texas A&M University System, College Station,
Texas 77843.
NJDEP. 2000. Revised Manual for New Jersey: Best Management Practices for Control of
Nonpoint Source Pollution from Stormwater.
http://www.state.nj.us/dep/watershedmgt/bmpmanual.html
6-15
Novotny, V. and G. Chesters. 1981. Handbook of Nonpoint Pollution Sources and
Management. Van Nostrand and Reinhold Environmental Engineering Series, New
York, NY.
Sample, D.J., J.P. Heaney, L.T. Wright, C.Y. Fan, F.H. Lai, and R. Field. 2003. Cost of Best
Management Practices and Associated Land for Urban Stormwater Control. Journal of
Water Resources Planning and Management, Vol. 129, No. 1, pp. 59-68.
Schueler, T.R. 1987. Controlling Urban Runoff: A Practical Manual for Planning and Designing
Urban BMPs. Metropolitan Washington Council of Governments, Washington, DC.
Southeastern Wisconsin Regional Planning Commission (SWRPC). 1991. “Costs of Urban
Nonpoint Source Water Pollution Control Measures”. Technical Report No. 31
Waukesha, WI, 107 pp.
U.S. Department of Defense (DOD). 1995. Parametric Cost Estimating Handbook. Joint
Government/Industry Initiative.
U.S. EPA. 1990. “National Pollutant Discharge Elimination System Permit Application
Regulations for Storm Water Discharges.” (Phase I Final Rule). Federal Register 55 FR
47990.
U.S. EPA. 1995. Economic Benefits of Runoff Controls. EPA-841-S-95-002. Office of
Wetlands, Oceans, and Watersheds, Washington, DC.
U.S. EPA. 1998. National Pollutant Discharge Elimination System–Proposed Regulations for
Revision of the Water Pollution Control Program Addressing Storm Water Discharges.
(Phase II Proposed Rule). Federal Register. 63FR1536, January 8.
U.S. EPA. 1999. Preliminary Data Summary of Urban Storm Water Best Management
Practices. EPA-821-R-99-012. August. Office of Water, Washington, DC.
U.S. EPA. 2000. A Guide to Developing and Documenting Cost Estimates During the
Feasibility Study. EPA/540/R-00/002. OSWER 9355.0-75. July.
U.S. EPA. 2003. National menu of best management practices for storm water Phase II. Office
of Water, Washington, DC.
http://cfpub.epa.gov/npdes/stormwater/menuofbmps/menu.cfm.
WERF. 2003. Post-Project Monitoring of BMPs/SUDS to Determine Performance and WholeLife Costs. Project Number: 01-CTS-21T.
Wiegand, C., T. Schueler, W. Chittenden, and D. Jellick. 1986. Cost of Urban Runoff Quality
Controls. pp. 366-380. In: Urban Runoff Quality. Engineering Foundation Conference.
ASCE, Henniker, NH. June 23-27.
Young, G.K., S. Stein, P. Cole, T. Kammer, F. Graziano, and F. Bank. 1996. Evaluation and
Management of Highway Runoff. Water Quality Technical Report. Department of
Environmental Programs, Metropolitan Washington Council of Governments,
Washington, DC.
6-16
@
Figure 6-1. Regional Rainfall Zones for the United States
Rainfall Zones can be found in the NPDES Phase I regulations, 40 CFR Part 122, Appendix E
(U.S. EPA, 1990)
6-17
Download